(mips_elf_make_msymbol_special): Don't set the low bit in the
symbol's address.
(mips_read_pc, mips_unwind_pc, mips_addr_bits_remove): Strip bit
indicating mips16 address, if present.
(mips_write_pc): Set bit indicating mips16 address when in a mips16
function.
(mips_eabi_push_dummy_call, mips_o64_push_dummy_call): Likewise,
but for each function pointer argument to inferior function call.
(mips_convert_register_gpreg_case_p): New functions.
(mips_convert_register_p): Invoke new functions above.
(mips_register_to_value): Add case for fetching value shorter
than 64 bits from a 64-bit register.
(mips_value_to_register): Add case for storing value shorter
than 64 bits into a 64-bit register.
cooked values that are being transferred to 64-bit raw registers.
(mips_pseudo_register_read): Revise to preserve symmetry with
mips_pseudo_register_write().
* valops.c (value_one): Use get_array_bounds to compute the number
of array elements instead of dividing the length of the array by the
length of the element types.
* valarith.c (value_complement, value_neg): Likewise.
This is a nasty interaction between Python and GDB. Basically,
Python causes some macros to be unilaterally defined in order
to turns some features on:
/* Define to activate features from IEEE Stds 1003.1-2001 */
#define _POSIX_C_SOURCE 200112L
/* Define to the level of X/Open that your system supports */
#define _XOPEN_SOURCE 600
But the problem is that they turn off defines provided by some
system headers on which we depend. Namely:
* sys/siginfo.h:
#if _SGIAPI
#define siginfo __siginfo
#endif
* sys/ucontext.h:
#if _SGIAPI && !defined(__SGI_NOUCONTEXT_COMPAT)
[...]
#define fp_r __fp_r
[...]
#define fp_csr __fp_csr
[...]
#endif
The important macro here is _SGIAPI, defined as follow in standards.h:
#define _SGIAPI ((defined(_SGI_SOURCE) && \
_NO_POSIX && _NO_XOPEN4 && _NO_XOPEN5) || \
(_ANSIMODE && _NO_POSIX && _NO_XOPEN4 && _NO_XOPEN5))
If one builds GDB without Python, then _SGIAPI is true, and all is fine.
But building with Python causes both _POSIX_C_SOURCE and _XOPEN_SOURCE
to trip all the _NO_[...] tests (_NO_POSIX, _NO_XOPEN4, _NO_XOPEN5).
And so we get build failures because we try to use undefined types, or
non-existent component names inside the regset structure.
The latter problem is observed only within irix5-nat.c, which means
that it is specific to IRIX. So it's easy to write the code in a way
that it does not require the macros (just use the real component names,
rather than relying on the macros to do the translation).
The former, on the other hand, is a little trickier, because the problem
occurs inside a generic unit (procfs.c). The solution I chose was to
adjust the configure script to add -Dsiginfo=__siginfo to the CPPFLAGS
if building with python using GCC on IRIX.
We hadn't seen this sort of issue up to now because the affect units
have not been dependent on the python includes up to now. Recent changes
have made them indirectly dependent on Python, thus triggering the issues.
gdb/ChangeLog:
* irix5-nat.c: Replace fp_r, fp_regs and fp_scr by __fp_r, __fp_regs
and __fp_scr respectively throughout.
* configure.ac: Compile with -Dsiginfo=__siginfo if building with
Python using GCC on IRIX.
* configure: Regenerate.
(dwarf2_per_cu_quick_data): Move definition down to "quick functions"
section. Delete members lines full_names. Rename read_lines to
no_file_data.
(quick_file_names): New struct.
(hash_file_name_entry, eq_file_name_entry): New functions.
(delete_file_name_entry, create_quick_file_names_table): New functions.
(dwarf2_read_index): Initialize quick_file_names_table.
(dw2_get_file_names): Renamed from dw2_require_line_header.
Return pointer to quick_file_names entry. All callers updated.
(dw2_get_real_path): Renamed from dw2_require_full_path.
Replace per_cu arg with pointer to quick_file_names entry.
All callers updated.
(dw2_free_cached_file_names): New function.
(dw2_forget_cached_source_info): Call it (via htab_traverse_noresize).
(dwarf2_initialize_objfile): Initialize quick_file_names_table in
the OBJF_READNOW case.
(dwarf2_free_objfile): Rewrite freeing of quick file names tables.
* dwarf2read.c (struct psymtab_cu_index_map): New struct.
(hash_psymtab_cu_index, eq_psymtab_cu_index): New functions.
(struct addrmap_index_data): New struct.
(add_address_entry): Remove arg `pst', new args `start', `end'.
(add_address_entry_worker, write_address_map): New functions.
(write_psymtabs_to_index): Address table generation moved to
write_address_map. Build a table mapping psymtab to CU index
to pass to it.
C++ class type is returned, fix incorrect enclosing type / embedded
offset. If internal variable is returned, allocate new internalvar
value using value_of_internalvar.
* NEWS: Document changes in behavior of "print x = 0" and similar
expressions.
Fix step_resume_breakpoint unsaved during an infcall.
* gdbthread.h (struct thread_control_state): Move here field
step_resume_breakpoint ...
(struct thread_info): ... from here.
* infrun.c (save_infcall_control_state): Reset
control.step_resume_breakpoint to NULL.
(restore_infcall_control_state, discard_infcall_control_state): Delete
control.step_resume_breakpoint.
* arm-linux-tdep.c, infrun.c, thread.c: Update all the references to
the moved field.
gdb/testsuite/
Fix step_resume_breakpoint unsaved during an infcall.
* gdb.base/step-resume-infcall.exp: New file.
* gdb.base/step-resume-infcall.c: New file.
Rename and move inferior_thread_state and inferior_status.
* gdbthread.h (struct thread_control_state): New struct, move fields
step_range_start, step_range_end, step_frame_id, step_stack_frame_id,
trap_expected, proceed_to_finish, in_infcall, step_over_calls,
stop_step and stop_bpstat here from struct thread_info.
(struct thread_suspend_state): New struct, move field stop_signal here
from struct thread_info.
(struct thread_info): Move the fields above from this struct.
* inferior.h: Move the inferior_thread_state and inferior_status
declarations comment to their definitions at infrun.c.
(struct inferior_control_state): New struct, move field stop_soon from
struct inferior here.
(struct inferior_suspend_state): New empty struct.
(struct inferior): New fields control and suspend. Move out field
stop_soon.
* infrun.c (struct inferior_thread_state): Rename to ...
(infcall_suspend_state): ... here. Replace field stop_signal by
fields thread_suspend and inferior_suspend.
(save_inferior_thread_state): Rename to ...
(save_infcall_suspend_state): ... here. New variable inf. Update the
code for new fields.
(restore_inferior_thread_state): Rename to ...
(restore_infcall_suspend_state): ... here. New variable inf. Update
the code for new fields.
(do_restore_inferior_thread_state_cleanup): Rename to ...
(do_restore_infcall_suspend_state_cleanup): ... here.
(make_cleanup_restore_inferior_thread_state): Rename to ...
(make_cleanup_restore_infcall_suspend_state): ... here.
(discard_inferior_thread_state): Rename to ...
(discard_infcall_suspend_state): ... here.
(get_inferior_thread_state_regcache): Rename to ...
(get_infcall_suspend_state_regcache): ... here.
(struct inferior_status): Rename to ...
(struct infcall_control_state): ... here. Replace fields
step_range_start, step_range_end, step_frame_id, step_stack_frame_id,
trap_expected, proceed_to_finish, in_infcall, step_over_calls,
stop_step, stop_bpstat and stop_soon by fields thread_control and
inferior_control.
(save_inferior_status): Rename to ...
(save_infcall_control_state): ... here. Update the code for new
fields.
(restore_inferior_status): Rename to ...
(restore_infcall_control_state): ... here. Update the code for new
fields.
(do_restore_inferior_status_cleanup): Rename to ...
(do_restore_infcall_control_state_cleanup): ... here.
(make_cleanup_restore_inferior_status): Rename to ...
(make_cleanup_restore_infcall_control_state): ... here.
(discard_inferior_status): Rename to ...
(discard_infcall_control_state): ... here.
* alpha-tdep.c, breakpoint.c, dummy-frame.c, dummy-frame.h,
exceptions.c, fbsd-nat.c, gdbthread.h, infcall.c, infcmd.c,
inferior.c, inferior.h, infrun.c, linux-nat.c, mi/mi-interp.c,
mips-tdep.c, procfs.c, solib-irix.c, solib-osf.c, solib-spu.c,
solib-sunos.c, solib-svr4.c, thread.c, windows-nat.c: Update all the
references to the moved fields and renamed functions.
bfd_uses_spe_extensions should only be used when BFD has been built
with ELF support. The typical way of checking that in GDB is to use
the HAVE_ELF macro.
gdb/ChangeLog:
* rs6000-tdep.c (bfd_uses_spe_extensions): Use bfd_elf_get_obj_attr_int
only if HAVE_ELF is defined.
* gdb.asm/asm-source.exp: Add new target "powerpc64"; expect leading `.'
on ppc64's symbols; "finish" and "return" commands must accept to
show the caller line again as well as the line after.
* gdb.asm/powerpc64.inc: New file.
(gdb_expect_list): Add verbose -log call for each pattern.
* gdb.base/signals.exp (test_handle_all_print): Call it.
Reduce timeout increment from 6 minutes to 1 minute.
* gdb.server/ext-run.exp: Call it.
PR python/12212
* python/py-inferior.c (find_thread_object): Check if PIDGET
returns 0.
2010-11-23 Phil Muldoon <pmuldoon@redhat.com>
PR python/12212
* gdb.python/python.exp: Check that selected_thread raises an
error when no inferior is loaded.
A recent change in check_typedef caused the following regression,
considering:
type String_Access is access String;
S1 : String_Access := null;
Trying to print S1, we get:
(gdb) print s1
$1 = (string) (null)
The type name is wrong. We were expecting:
(gdb) print s1
$1 = (string_bug.string_access) (null)
The extensive comment in this patch explains how pointers to arrays
are encoded when the array is a "fat pointer". What happened is
that the change in check_typedef broke a type-equality test that
we were performing. The test really was on the fringe, but it
turns out that, for what we're doing, we're really only interested
in the main-type portion of the type.
The patch adjust the check accordingly.
gdb/ChangeLog:
* ada-lang.c (ada_to_fixed_type): Expand function documentation.
Return the original type if the main type portions match rather
than when the type themselves match.
We have a variant record whose value is defined as follow:
type Discriminant_Record (Num1, Num2,
Num3, Num4 : Natural) is record
Field1 : My_Record_Array (1 .. Num2);
Field2 : My_Record_Array (Num1 .. 10);
Field3 : My_Record_Array (Num1 .. Num2);
Field4 : My_Record_Array (Num3 .. Num2);
Field5 : My_Record_Array (Num4 .. Num2);
end record;
Dire : Discriminant_Record (1, 7, 3, 0);
However, we're trying to print "Dire" before it is elaborated.
This is common if one breaks on a function and then starts doing
"info locals" for instance.
What happens is that GDB reads bogus values for fields Num1 to
Num4, and deduces a bogus (ginormouos) size for component
"Field1". The length is so large that it then later causes
an overflow in the overall record length computation. Things
go downhill from there, because length(field1) > length(record).
So, when after we've fetched the value of the record based on
the computed size, we crash trying to access unallocated memory
when accessing field1...
The first fix we can do is to check the size of the field
against the maximum object size. If it exceeds that size,
then we know the record will also exceed that size...
gdb/ChangeLog:
* ada-lang.c (ada_template_to_fixed_record_type_1):
For dynamic fields, check the field size against the maximum
object size.
On mips-irix, the debugger has trouble stepping over the following
line of code:
S: Shape'Class := R; <<<<---- STOP here
Here is what happens:
(gdb) n
warning: GDB can't find the start of the function at 0x7fff2bd8.
GDB is unable to find the start of the function at 0x7fff2bd8
and thus can't determine the size of that function's stack frame.
This means that GDB may be unable to access that stack frame, or
the frames below it.
This problem is most likely caused by an invalid program counter or
stack pointer.
However, if you think GDB should simply search farther back
from 0x7fff2bd8 for code which looks like the beginning of a
function, you can increase the range of the search using the `set
heuristic-fence-post' command.
0x7fff2bd8 in ?? ()
The program does in fact jump to this code location, which is a trampoline
located on the stack (there is an implicit call to a routine internally
generated by the Ada expander). As it is on the stack, GDB is naturally
unable to find the bounds of the current function, or any debugging
information, and is thus unable to continue.
This patch adds support for this sort of trampoline.
gdb/ChangeLog:
* mips-irix-tdep.c (mips_irix_n32_stack_tramp_frame_init): New
function.
(mips_irix_n32_stack_tramp_frame): New static global.
(mips_irix_init_abi): Add mips_irix_n32_stack_tramp_frame to
list of unwinder.
Type symbols that are used to decode the Ada task control block may be
duplicated in any unit that uses a tasking feature; an Ada lookup of
these symbols would collect all the duplicated instances and could
provoke the load of a huge amount of debug symbols in the process.
However, when decoding the task information, we are only interested in
one instance; and we have the full name of the symbol anyway. So use
a C lookup instead of an Ada lookup.
2010-11-05 Jerome Guitton <guitton@adacore.com>
* ada-tasks.c (get_tcb_types_info): Use C lookups to get
ATCB symbols.
When trying to resolve an incomplete type, if there is no complete
version of that type available, GDB can go in an infinite loop.
This is because ada_check_typedef makes a recursive call to itself,
in an attempt to make sure that the returned type is never a typedef.
However, when no complete type is found, the current logic causes us
to keep going indefinitely through the same path...
This patch fixes the problem by performing the recursive call to
ada_check_typedef only when a TYPE_CODE_TYPDEF layer needs to be
stripped.
gdb/ChangeLog:
* ada-lang.c (ada_check_typedef): Call ada_check_typedef only
if type1 is a typedef.
This is a new routine that is being generated by the expander in
the context of controlled types. This patch adds it to the list of such
routines so that the debugger does not show it when selecting the first
"user" frame (when switching tasks for instance).
gdb/ChangeLog:
* ada-lang.h (ADA_KNOWN_AUXILIARY_FUNCTION_NAME_PATTERNS): Add
"_finalizer" to the list.
gdb/ChangeLog:
* ada-typeprint.c (print_array_type): Fix formatting in comment.
(print_selected_record_field_types): Fix formatting in function
documentation.
A new-line was missing at the end of the output of the `show ravenscar
task-switching' command. For instance:
(gdb) show ravenscar task-switching
Support for Ravenscar task/thread switching is enabled(gdb)
This patch fixes it.
gdb/ChangeLog:
* ravenscar-thread.c (show_ravenscar_task_switching_command):
Add missing '\n' in output.
The GNAT Ravenscar implementation has recently been enhanced to allow
programs using that runtime to run on multi-cpu systems. One of the
changes that were made is that the name of the symbol we use to
determine which task is the currently-running task has changed.
This patch enhances the debugger to use the new name, and fallback on
the old name if not found (this is for compatibility with older versions
of the compiler/runtime).
gdb/ChangeLog:
* ravenscar-thread.c (running_thread_name): Change value.
(read_thread_id): Remove advance declaration.
(get_running_thread_msymbol): New function.
(has_ravenscar_runtime): Use get_running_thread_msymbol to
compute msym_running_thread.
(get_running_thread_id): Renames read_thread_id. Slight modifications
to not take any argument anymore, using get_running_thread_msymbol
to determine which symbol to use instead.
(ravenscar_running_thread): Use get_running_thread_id instead of
read_thread_id.