Move get_dyn_prop, currently a free function, to be a method on struct
type.
gdb/ChangeLog:
* gdbtypes.h (struct type) <get_dyn_prop>: New method.
(get_dyn_prop): Remove. Update all users to use
type::dyn_prop.
* gdbtypes.c (get_dyn_prop): Rename to...
(type::dyn_prop): ... this.
It is unused. The corresponding macro was removed in c3236f84c1 ("gdb:
remove TYPE_INCOMPLETE").
gdb/ChangeLog:
* gdbtypes.h (struct main_type) <flag_incomplete>: Remove.
The "HP platforms" comment prompted me to check if this was still used
somewhere. Apparently it's not, so remove it.
gdb/ChangeLog:
* gdbtypes.h (TYPE_INCOMPLETE): Remove.
* gdbtypes.c (recursive_dump_type): Remove use of
TYPE_INCOMPLETE.
In Ada, a type with variant parts can have a variable length. This
patch adds support for this to gdb, by integrating the length
computation into the dynamic type resolution code.
gdb/ChangeLog
2020-04-24 Tom Tromey <tromey@adacore.com>
* dwarf2/read.c (read_structure_type): Handle dynamic length.
* gdbtypes.c (is_dynamic_type_internal): Check
TYPE_HAS_DYNAMIC_LENGTH.
(resolve_dynamic_type_internal): Use TYPE_DYNAMIC_LENGTH.
* gdbtypes.h (TYPE_HAS_DYNAMIC_LENGTH, TYPE_DYNAMIC_LENGTH):
New macros.
(enum dynamic_prop_node_kind) <DYN_PROP_BYTE_SIZE>: New
constant.
gdb/testsuite/ChangeLog
2020-04-24 Tom Tromey <tromey@adacore.com>
* gdb.ada/variant.exp: New file
* gdb.ada/variant/pkg.adb: New file
* gdb.ada/variant/pck.adb: New file
This rewrites the existing variant part code to follow the new model
implemented in the previous patch. The old variant part code is
removed.
This only affects Rust for the moment. I tested this using various
version of the Rust compiler, including one that emits old-style enum
debuginfo, exercising the quirks code.
gdb/ChangeLog
2020-04-24 Tom Tromey <tromey@adacore.com>
* dwarf2/read.c (struct variant_field): Rewrite.
(struct variant_part_builder): New.
(struct nextfield): Remove "variant" field. Add "offset".
(struct field_info): Add "current_variant_part" and
"variant_parts".
(alloc_discriminant_info): Remove.
(alloc_rust_variant): New function.
(quirk_rust_enum): Update.
(dwarf2_add_field): Set "offset" member. Don't handle
DW_TAG_variant_part.
(offset_map_type): New typedef.
(convert_variant_range, create_one_variant)
(create_one_variant_part, create_variant_parts)
(add_variant_property): New functions.
(dwarf2_attach_fields_to_type): Call add_variant_property.
(read_structure_type): Don't handle DW_TAG_variant_part.
(handle_variant_part, handle_variant): New functions.
(handle_struct_member_die): Use them.
(process_structure_scope): Don't handle variant parts.
* gdbtypes.h (TYPE_FLAG_DISCRIMINATED_UNION): Remove.
(struct discriminant_info): Remove.
(enum dynamic_prop_node_kind) <DYN_PROP_DISCRIMINATED>: Remove.
(struct main_type) <flag_discriminated_union>: Remove.
* rust-lang.c (rust_enum_p, rust_empty_enum_p): Rewrite.
(rust_enum_variant): Return int. Remove "contents". Rewrite.
(rust_print_enum, rust_print_struct_def, rust_evaluate_subexp):
Update.
* valops.c (value_union_variant): Remove.
* value.h (value_union_variant): Don't declare.
When evaluating a DWARF expression, the dynamic type resolution code
will pass in a buffer of bytes via the property_addr_info. However,
the DWARF expression evaluator will then proceed to read memory from
the inferior, even when the request could be filled from this buffer.
This, in turn, is a problem in some cases; and specifically when
trying to handle the Ada scenario of extracting a variable-length
value from a packed array. Here, the ordinary DWARF expression cannot
be directly evaluated, because the data may appear at some arbitrary
bit offset. So, it is unpacked into a staging area and then the
expression is evaluated -- using an address of 0.
This patch fixes the problem by arranging for the DWARF evaluator, in
this case, to prefer passed-in memory when possible. The type of the
buffer in the property_addr_info is changed to an array_view so that
bounds checking can be done.
gdb/ChangeLog
2020-04-24 Tom Tromey <tromey@adacore.com>
* ada-lang.c (ada_discrete_type_high_bound, ada_discrete_type_low)
(ada_value_primitive_packed_val): Update.
* ada-valprint.c (ada_value_print_1): Update.
* dwarf2/loc.c (evaluate_for_locexpr_baton): New struct.
(dwarf2_locexpr_baton_eval): Take a property_addr_info rather than
just an address. Use evaluate_for_locexpr_baton.
(dwarf2_evaluate_property): Update.
* dwarf2/loc.h (struct property_addr_info) <valaddr>: Now an
array_view.
* findvar.c (default_read_var_value): Update.
* gdbtypes.c (compute_variant_fields_inner)
(resolve_dynamic_type_internal): Update.
(resolve_dynamic_type): Change type of valaddr parameter.
* gdbtypes.h (resolve_dynamic_type): Update.
* valarith.c (value_subscripted_rvalue): Update.
* value.c (value_from_contents_and_address): Update.
This patch adds the infrastructure for the new variant part code. At
this point, nothing uses this code. This is done in a separate patch
to make it simpler to review.
I examined a few possible approaches to handling variant parts. In
particular, I considered having a DWARF variant part be a union
(similar to how the Rust code works now); and I considered having type
fields have a flag indicating that they are variants.
Having separate types seemed bad conceptually, because these variants
aren't truly separate -- they rely on the "parent" type. And,
changing how fields worked seemed excessively invasive.
So, in the end I thought the approach taken in this patch was both
simple to implement and understand, without losing generality. The
idea in this patch is that all the fields of a type with variant parts
will be stored in a single field array, just as if they'd all been
listed directly. Then, the variants are attached as a dynamic
property. These control which fields end up in the type that's
constructed during dynamic type resolution.
gdb/ChangeLog
2020-04-24 Tom Tromey <tromey@adacore.com>
* gdbtypes.c (is_dynamic_type_internal): Check for variant parts.
(variant::matches, compute_variant_fields_recurse)
(compute_variant_fields_inner, compute_variant_fields): New
functions.
(resolve_dynamic_struct): Check for DYN_PROP_VARIANT_PARTS.
Use resolved_type after type is made.
(operator==): Add new cases.
* gdbtypes.h (TYPE_HAS_VARIANT_PARTS): New macro.
(struct discriminant_range, struct variant, struct variant_part):
New.
(union dynamic_prop_data) <variant_parts, original_type>: New
members.
(enum dynamic_prop_node_kind) <DYN_PROP_VARIANT_PARTS>: New constant.
(enum dynamic_prop_kind) <PROP_TYPE, PROP_VARIANT_PARTS>: New
constants.
* value.c (unpack_bits_as_long): Now public.
* value.h (unpack_bits_as_long): Declare.
This patch changes how complex types are created. init_complex_type
and arch_complex_type are unified, and complex types are reused, by
attaching them to the underlying scalar type.
gdb/ChangeLog
2020-04-01 Tom Tromey <tom@tromey.com>
* stabsread.c (rs6000_builtin_type, read_sun_floating_type)
(read_range_type): Update.
* mdebugread.c (basic_type): Update.
* go-lang.c (build_go_types): Use init_complex_type.
* gdbtypes.h (struct main_type) <complex_type>: New member.
(init_complex_type): Update.
(arch_complex_type): Don't declare.
* gdbtypes.c (init_complex_type): Remove "objfile" parameter.
Make name if none given. Use alloc_type_copy. Look for cached
complex type.
(arch_complex_type): Remove.
(gdbtypes_post_init): Use init_complex_type.
* f-lang.c (build_fortran_types): Use init_complex_type.
* dwarf2/read.c (read_base_type): Update.
* d-lang.c (build_d_types): Use init_complex_type.
* ctfread.c (read_base_type): Update.
While working on complex number support, I found a couple of
apparently obsolete coments. This removes them.
2020-03-07 Tom Tromey <tom@tromey.com>
* valops.c (value_literal_complex): Remove obsolete comment.
* gdbtypes.h (enum type_code) <TYPE_CODE_FLT>: Remove obsolete
comment.
This is a refactoring. Instead of a plain unsigned value, use an enum
bitfield.
gdb/ChangeLog:
2019-12-20 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* dwarf2read.c (is_valid_DW_AT_calling_convention_for_subroutine):
New function.
(read_subroutine_type): Validate the parsed
DW_AT_calling_convention value before assigning it to a
subroutine's calling_convention attribute.
* gdbtypes.h (struct func_type) <calling_convention>: Use
an enum bitfield as its type, instead of plain unsigned.
Change-Id: Ibc6b2f71e885cbc5c3c9d49734f7125acbfd1bcd
Extend GDB's internal representation of types to include the
DW_AT_calling_convention, DW_AT_defaulted, and DW_AT_deleted attributes
that were introduced in DWARF5.
These attributes will be helpful in a future patch about infcall'ing
functions with call-by-value parameters. GDB will use the attributes
to decide whether the type of a call-by-value parameter is implicitly
pass-by-reference.
gdb/ChangeLog:
2019-12-20 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* dwarf2read.c (dwarf2_add_member_fn): Read the DW_AT_defaulted
and DW_AT_deleted attributes of a function.
(read_structure_type): Read the DW_AT_calling_convention attribute
of a type.
(is_valid_DW_AT_defaulted): New function.
(is_valid_DW_AT_calling_convention_for_type): New function.
* gdbtypes.h: Include dwarf2.h.
(struct fn_field)<defaulted>: New field to store the
DW_AT_defaulted attribute.
(struct fn_field)<is_deleted>: New field to store the
DW_AT_deleted attribute.
(struct cplus_struct_type)<calling_convention>: New field to store
the DW_AT_calling_convention attribute.
(TYPE_FN_FIELD_DEFAULTED): New macro.
(TYPE_FN_FIELD_DELETED): New macro.
(TYPE_CPLUS_CALLING_CONVENTION): New macro.
* gdbtypes.c (dump_fn_fieldlists): Update for the changes made
to the .h file.
(print_cplus_stuff): Likewise.
Change-Id: I54192f363115b78ec7435a8563b73fcace420765
The overload resolution mechanism assigns badness values to the
necessary conversions to be made on types to pick a champion. A
badness value consists of a "rank" that scores the conversion and a
"subrank" to differentiate conversions of the same kind.
An auxiliary function, 'sum_ranks', is used for adding two badness
values. In all of its uses, except two, 'sum_ranks' is used for
populating the subrank of a badness value. The two exceptions are in
'rank_one_type':
~~~
/* See through references, since we can almost make non-references
references. */
if (TYPE_IS_REFERENCE (arg))
return (sum_ranks (rank_one_type (parm, TYPE_TARGET_TYPE (arg), NULL),
REFERENCE_CONVERSION_BADNESS));
if (TYPE_IS_REFERENCE (parm))
return (sum_ranks (rank_one_type (TYPE_TARGET_TYPE (parm), arg, NULL),
REFERENCE_CONVERSION_BADNESS));
~~~
Here, the result of a recursive call is combined with
REFERENCE_CONVERSION_BADNESS. This leads to the problem of
over-punishment by combining two ranks. Consider this:
void an_overloaded_function (const foo &);
void an_overloaded_function (const foo &&);
...
foo arg;
an_overloaded_function(arg);
When ranking 'an_overloaded_function (const foo &)', the badness
values REFERENCE_CONVERSION_BADNESS and CV_CONVERSION_BADNESS are
combined, whereas 'rank_one_type' assigns only the
REFERENCE_CONVERSION_BADNESS value to 'an_overloaded_function (const
foo &&)' (there is a different execution flow for that). This yields
in GDB picking the latter function as the overload champion instead of
the former.
In fact, the 'rank_one_type' function should have given
'an_overloaded_function (const foo &)' the CV_CONVERSION_BADNESS
value, with the see-through referencing increasing the subrank a
little bit. This can be achieved by introducing a new badness value,
REFERENCE_SEE_THROUGH_BADNESS, which bumps up the subrank only, and
using it in the two "exceptional" cases of 'sum_ranks'.
gdb/ChangeLog:
2019-12-06 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* gdbtypes.h: Define the REFERENCE_SEE_THROUGH_BADNESS value.
* gdbtypes.c (rank_one_type): Use REFERENCE_SEE_THROUGH_BADNESS
for ranking see-through reference cases.
gdb/testsuite/ChangeLog:
2019-12-06 Tankut Baris Aktemur <tankut.baris.aktemur@intel.com>
* gdb.cp/rvalue-ref-overload.cc: Add a case that involves both
CV and reference conversion for overload resolution.
* gdb.cp/rvalue-ref-overload.exp: Test it.
Change-Id: I39ae6505ab85ad0bd21915368c82540ceeb3aae9
From what I can tell, set_gdbarch_bits_big_endian has never been used.
That is, all architectures since its introduction have simply used the
default, which is simply check the architecture's byte-endianness.
Because this interferes with the scalar_storage_order code, this patch
removes this gdbarch setting entirely. In some places,
type_byte_order is used rather than the plain gdbarch.
gdb/ChangeLog
2019-12-04 Tom Tromey <tromey@adacore.com>
* ada-lang.c (decode_constrained_packed_array)
(ada_value_assign, value_assign_to_component): Update.
* dwarf2loc.c (rw_pieced_value, access_memory)
(dwarf2_compile_expr_to_ax): Update.
* dwarf2read.c (dwarf2_add_field): Update.
* eval.c (evaluate_subexp_standard): Update.
* gdbarch.c, gdbarch.h: Rebuild.
* gdbarch.sh (bits_big_endian): Remove.
* gdbtypes.h (union field_location): Update comment.
* target-descriptions.c (make_gdb_type): Update.
* valarith.c (value_bit_index): Update.
* value.c (struct value) <bitpos>: Update comment.
(unpack_bits_as_long, modify_field): Update.
* value.h (value_bitpos): Update comment.
Change-Id: I379b5e0c408ec8742f7a6c6b721108e73ed1b018
Testing the scalar_storage_order patch pointed out that it does not
handle floating point properly. This patch fixes this problem.
gdb/ChangeLog
2019-12-04 Tom Tromey <tromey@adacore.com>
* dwarf2read.c (dwarf2_init_float_type)
(dwarf2_init_complex_target_type): Add byte_order parameter.
(read_base_type): Compute byte order earlier.
* gdbtypes.c (init_float_type): Add byte_order parameter.
* gdbtypes.h (init_float_type): Add byte_order parameter.
gdb/testsuite/ChangeLog
2019-12-04 Tom Tromey <tromey@adacore.com>
* gdb.base/endianity.c (struct otherendian) <f>: New field.
(main): Initialize it.
* gdb.base/endianity.exp: Update.
Change-Id: Ic02eb711d80ce678ef0ecf8c506a626e441b8440
Currently GDB supports a byte or bit stride on arrays, in DWARF this
would be DW_AT_bit_stride or DW_AT_byte_stride on DW_TAG_array_type.
However, DWARF can also support DW_AT_byte_stride or DW_AT_bit_stride
on DW_TAG_subrange_type, the tag used to describe each dimension of an
array.
Strides on subranges are used by gFortran to represent Fortran arrays,
and this commit adds support for this to GDB.
I've extended the range_bounds struct to include the stride
information. The name is possibly a little inaccurate now, but this
still sort of makes sense, the structure represents information about
the bounds of the range, and also how to move from the lower to the
upper bound (the stride).
I've added initial support for bit strides, but I've never actually
seen an example of this being generated. Further, I don't really see
right now how GDB would currently handle a bit stride that was not a
multiple of the byte size as the code in, for example,
valarith.c:value_subscripted_rvalue seems geared around byte
addressing. As a consequence if we see a bit stride that is not a
multiple of 8 then GDB will give an error.
gdb/ChangeLog:
* dwarf2read.c (read_subrange_type): Read bit and byte stride and
create a range with stride where appropriate.
* f-valprint.c: Include 'gdbarch.h'.
(f77_print_array_1): Take the stride into account when walking the
array. Also convert the stride into addressable units.
* gdbtypes.c (create_range_type): Initialise the stride to
constant zero.
(create_range_type_with_stride): New function, initialise the
range as normal, and then setup the stride.
(has_static_range): Include the stride here. Also change the
return type to bool.
(create_array_type_with_stride): Consider the range stride if the
array isn't given its own stride.
(resolve_dynamic_range): Resolve the stride if needed.
* gdbtypes.h (struct range_bounds) <stride>: New member variable.
(struct range_bounds) <flag_is_byte_stride>: New member variable.
(TYPE_BIT_STRIDE): Define.
(TYPE_ARRAY_BIT_STRIDE): Define.
(create_range_type_with_stride): Declare.
* valarith.c (value_subscripted_rvalue): Take range stride into
account when walking the array.
gdb/testsuite/ChangeLog:
* gdb.fortran/derived-type-striding.exp: New file.
* gdb.fortran/derived-type-striding.f90: New file.
* gdb.fortran/array-slices.exp: New file.
* gdb.fortran/array-slices.f90: New file.
Change-Id: I9af2bcd1f2d4c56f76f5f3f9f89d8f06bef10d9a
While debugging gdb, I noticed that the bitfields in a range_bounds
were signed, causing the values of these fields to be -1.
I think this is odd; and while we haven't yet committed to boolean
bitfields, I think it is a small improvement to change these types to
unsigned.
gdb/ChangeLog
2019-11-28 Tom Tromey <tom@tromey.com>
* gdbtypes.h (struct range_bounds) <flag_upper_bound_is_count,
flag_bound_evaluated>: Now unsigned.
Change-Id: Ia377fd931594bbf8653180d4dcb4e60354d90139
- Rationale:
It is possible for compilers to indicate the desired byte order
interpretation of scalar variables using the DWARF attribute:
DW_AT_endianity
A type flagged with this variable would typically use one of:
DW_END_big
DW_END_little
which instructs the debugger what the desired byte order interpretation
of the variable should be.
The GCC compiler (as of V6) has a mechanism for setting the desired byte
ordering of the fields within a structure or union. For, example, on a
little endian target, a structure declared as:
struct big {
int v;
short a[4];
} __attribute__( ( scalar_storage_order( "big-endian" ) ) );
could be used to ensure all the structure members have a big-endian
interpretation (the compiler would automatically insert byte swap
instructions before and after respective store and load instructions).
- To reproduce
GCC V8 is required to correctly emit DW_AT_endianity DWARF attributes
in all situations when the scalar_storage_order attribute is used.
A fix for (dwarf endianity instrumentation) for GCC V6-V7 can be found
in the URL field of the following PR:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82509
- Test-case:
A new test case (testsuite/gdb.base/endianity.*) is included with this
patch.
Manual testing for mixed endianity code has also been done with GCC V8.
See:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82509#c4
- Observed vs. expected:
Without this change, using scalar_storage_order that doesn't match the
target, such as
struct otherendian
{
int v;
} __attribute__( ( scalar_storage_order( "big-endian" ) ) );
would behave like the following on a little endian target:
Breakpoint 1 at 0x401135: file endianity.c, line 41.
(gdb) run
Starting program: /home/pjoot/freeware/t/a.out
Missing separate debuginfos, use: debuginfo-install glibc-2.17-292.el7.x86_64
Breakpoint 1, main () at endianity.c:41
41 struct otherendian o = {3};
(gdb) n
43 do_nothing (&o); /* START */
(gdb) p o
$1 = {v = 50331648}
(gdb) p /x
$2 = {v = 0x3000000}
whereas with this gdb enhancement we can access the variable with the user
specified endianity:
Breakpoint 1, main () at endianity.c:41
41 struct otherendian o = {3};
(gdb) p o
$1 = {v = 0}
(gdb) n
43 do_nothing (&o); /* START */
(gdb) p o
$2 = {v = 3}
(gdb) p o.v = 4
$3 = 4
(gdb) p o.v
$4 = 4
(gdb) x/4xb &o.v
0x7fffffffd90c: 0x00 0x00 0x00 0x04
(observe that the 4 byte int variable has a big endian representation in the
hex dump.)
gdb/ChangeLog
2019-11-21 Peeter Joot <peeter.joot@lzlabs.com>
Byte reverse display of variables with DW_END_big, DW_END_little
(DW_AT_endianity) dwarf attributes if different than the native
byte order.
* ada-lang.c (ada_value_binop):
Use type_byte_order instead of gdbarch_byte_order.
* ada-valprint.c (printstr):
(ada_val_print_string):
* ada-lang.c (value_pointer):
(ada_value_binop):
Use type_byte_order instead of gdbarch_byte_order.
* c-lang.c (c_get_string):
Use type_byte_order instead of gdbarch_byte_order.
* c-valprint.c (c_val_print_array):
Use type_byte_order instead of gdbarch_byte_order.
* cp-valprint.c (cp_print_class_member):
Use type_byte_order instead of gdbarch_byte_order.
* dwarf2loc.c (rw_pieced_value):
Use type_byte_order instead of gdbarch_byte_order.
* dwarf2read.c (read_base_type): Handle DW_END_big,
DW_END_little
* f-lang.c (f_get_encoding):
Use type_byte_order instead of gdbarch_byte_order.
* findvar.c (default_read_var_value):
Use type_byte_order instead of gdbarch_byte_order.
* gdbtypes.c (check_types_equal):
Require matching TYPE_ENDIANITY_NOT_DEFAULT if set.
(recursive_dump_type): Print TYPE_ENDIANITY_BIG,
and TYPE_ENDIANITY_LITTLE if set.
(type_byte_order): new function.
* gdbtypes.h (TYPE_ENDIANITY_NOT_DEFAULT): New macro.
(struct main_type) <flag_endianity_not_default>:
New field.
(type_byte_order): New function.
* infcmd.c (default_print_one_register_info):
Use type_byte_order instead of gdbarch_byte_order.
* p-lang.c (pascal_printstr):
Use type_byte_order instead of gdbarch_byte_order.
* p-valprint.c (pascal_val_print):
Use type_byte_order instead of gdbarch_byte_order.
* printcmd.c (print_scalar_formatted):
Use type_byte_order instead of gdbarch_byte_order.
* solib-darwin.c (darwin_current_sos):
Use type_byte_order instead of gdbarch_byte_order.
* solib-svr4.c (solib_svr4_r_ldsomap):
Use type_byte_order instead of gdbarch_byte_order.
* stap-probe.c (stap_modify_semaphore):
Use type_byte_order instead of gdbarch_byte_order.
* target-float.c (target_float_same_format_p):
Use type_byte_order instead of gdbarch_byte_order.
* valarith.c (scalar_binop):
(value_bit_index):
Use type_byte_order instead of gdbarch_byte_order.
* valops.c (value_cast):
Use type_byte_order instead of gdbarch_byte_order.
* valprint.c (generic_emit_char):
(generic_printstr):
(val_print_string):
Use type_byte_order instead of gdbarch_byte_order.
* value.c (unpack_long):
(unpack_bits_as_long):
(unpack_value_bitfield):
(modify_field):
(pack_long):
(pack_unsigned_long):
Use type_byte_order instead of gdbarch_byte_order.
* findvar.c (unsigned_pointer_to_address):
(signed_pointer_to_address):
(unsigned_address_to_pointer):
(address_to_signed_pointer):
(default_read_var_value):
(default_value_from_register):
Use type_byte_order instead of gdbarch_byte_order.
* gnu-v3-abi.c (gnuv3_make_method_ptr):
Use type_byte_order instead of gdbarch_byte_order.
* riscv-tdep.c (riscv_print_one_register_info):
Use type_byte_order instead of gdbarch_byte_order.
gdb/testsuite/ChangeLog
2019-11-21 Peeter Joot <peeter.joot@lzlabs.com>
* gdb.base/endianity.c: New test.
* gdb.base/endianity.exp: New file.
Change-Id: I4bd98c1b4508c2d7c5a5dbb15d7b7b1cb4e667e2
In Ada, the programmer can request that a range type with a non-zero
base be stored in the minimal number of bits required for the range.
This is done by biasing the values; so, for example, a range of -7..-4
may be stored as two bits with a bias of -7.
This patch implements this for gdb. It is done by adding a bias to
struct range_bounds and then adjusting a few spots to handle this.
The test case is written to use -fgnat-encodings=minimal, but a future
compiler patch will change the compiler to emit DW_AT_GNU_bias with
-fgnat-encodings=gdb. It seemed good to get the gdb patch in first.
Tested on x86-64 Fedora 29; plus a variety of targets using AdaCore's
internal test suite.
gdb/ChangeLog
2019-09-03 Tom Tromey <tromey@adacore.com>
* ada-valprint.c (ada_val_print_num): Don't recurse for range
types.
(has_negatives): Unbias a range type bound.
* dwarf2read.c (read_subrange_type): Handle DW_AT_GNU_bias.
* gdbtypes.c (operator==): Handle new field.
(create_range_type): Add "bias" parameter.
(create_static_range_type, resolve_dynamic_range): Update.
* gdbtypes.h (struct range_bounds) <bias>: New member.
(create_range_type): Add bias parameter.
* printcmd.c (print_scalar_formatted): Unbias range types.
* value.c (unpack_long): Unbias range types.
(pack_long): Bias range types.
gdb/testsuite/ChangeLog
2019-09-03 Tom Tromey <tromey@adacore.com>
* gdb.ada/bias.exp: New file.
* gdb.ada/bias/bias.adb: New file.
* gdb.ada/print_chars.exp: Add regression test.
* gdb.ada/print_chars/foo.adb (My_Character): New type.
(MC): New variable.
I touched symtab.h and was surprised to see how many files were
rebuilt. I looked into it a bit, and found that defs.h includes
gdbarch.h, which in turn includes many things.
gdbarch.h is only needed by a minority ofthe files in gdb, so this
patch removes the include from defs.h and updates the fallout.
I did "wc -l" on the files in build/gdb/.deps; this patch reduces the
line count from 139935 to 137030; so there are definitely future
build-time savings here.
Note that while I configured with --enable-targets=all, it's possible
that some *-nat.c file needs an update. I could not test all of
these. The buildbot caught a few problems along these lines.
gdb/ChangeLog
2019-07-10 Tom Tromey <tom@tromey.com>
* defs.h: Don't include gdbarch.h.
* aarch64-ravenscar-thread.c, aarch64-tdep.c, alpha-bsd-tdep.h,
alpha-linux-tdep.c, alpha-mdebug-tdep.c, arch-utils.h, arm-tdep.h,
ax-general.c, btrace.c, buildsym-legacy.c, buildsym.h, c-lang.c,
cli/cli-decode.h, cli/cli-dump.c, cli/cli-script.h,
cli/cli-style.h, coff-pe-read.h, compile/compile-c-support.c,
compile/compile-cplus.h, compile/compile-loc2c.c, corefile.c,
cp-valprint.c, cris-linux-tdep.c, ctf.c, d-lang.c, d-namespace.c,
dcache.c, dicos-tdep.c, dictionary.c, disasm-selftests.c,
dummy-frame.c, dummy-frame.h, dwarf2-frame-tailcall.c,
dwarf2expr.c, expression.h, f-lang.c, frame-base.c,
frame-unwind.c, frv-linux-tdep.c, gdbarch-selftests.c, gdbtypes.h,
go-lang.c, hppa-nbsd-tdep.c, hppa-obsd-tdep.c, i386-dicos-tdep.c,
i386-tdep.h, ia64-vms-tdep.c, interps.h, language.c,
linux-record.c, location.h, m2-lang.c, m32r-linux-tdep.c,
mem-break.c, memattr.c, mn10300-linux-tdep.c, nios2-linux-tdep.c,
objfiles.h, opencl-lang.c, or1k-linux-tdep.c, p-lang.c,
parser-defs.h, ppc-tdep.h, probe.h, python/py-record-btrace.c,
record-btrace.c, record.h, regcache-dump.c, regcache.h,
riscv-fbsd-tdep.c, riscv-linux-tdep.c, rust-exp.y,
sh-linux-tdep.c, sh-nbsd-tdep.c, source-cache.c,
sparc-nbsd-tdep.c, sparc-obsd-tdep.c, sparc-ravenscar-thread.c,
sparc64-fbsd-tdep.c, std-regs.c, target-descriptions.h,
target-float.c, tic6x-linux-tdep.c, tilegx-linux-tdep.c, top.c,
tracefile.c, trad-frame.c, type-stack.h, ui-style.c, utils.c,
utils.h, valarith.c, valprint.c, varobj.c, x86-tdep.c,
xml-support.h, xtensa-linux-tdep.c, cli/cli-cmds.h: Update.
* s390-linux-nat.c, procfs.c, inf-ptrace.c: Likewise.
This fixes a couple of comments in gdbtypes.h. One comment had a
typo; and another comment referred to "Moto", which is presumably some
long-gone Motorola-related project.
Tested by rebuilding.
gdb/ChangeLog
2019-05-30 Tom Tromey <tromey@adacore.com>
* gdbtypes.h (struct range_bounds) <flag_upper_bound_is_count>:
Fix comment.
(TYPE_ARRAY_UPPER_BOUND_IS_UNDEFINED): Rewrite comment.
Types in Fortran can have the 'allocatable' qualifier attached to
indicate that memory needs to be explicitly allocated by the user.
This patch extends GDB to show this qualifier when printing types.
Lots of tests results are then updated to include this new qualifier
in the expected results.
gdb/ChangeLog:
* f-typeprint.c (f_type_print_base): Print 'allocatable' type
qualifier.
* gdbtypes.h (TYPE_IS_ALLOCATABLE): Define.
gdb/testsuite/ChangeLog:
* gdb.fortran/vla-datatypes.exp: Update expected results.
* gdb.fortran/vla-ptype.exp: Likewise.
* gdb.fortran/vla-type.exp: Likewise.
* gdb.fortran/vla-value.exp: Likewise.
ada-lang.c is a bit too eager trying to decode unions in the Ada style
-- looking for discriminants and such. This causes crashes when
printing a non-Ada union in Ada mode, something that can easily happen
when printing a value from history or certain registers on AArch64.
This patch fixes the bug by changing ada-lang.c to only apply special
Ada treatment to types coming from an Ada CU. This in turn required a
couple of surprising changes.
First, some of the Ada code was already using HAVE_GNAT_AUX_INFO to
decide whether a type had already been fixed -- such types had
INIT_CPLUS_SPECIFIC called on them. This patch changes these spots to
use the "none" identifier instead.
This then required changing value_rtti_type to avoid changing the
language-specific object attached to an Ada type, which seems like a
good change regardless.
Tested on x86-64 Fedora 29.
gdb/ChangeLog
2019-04-19 Tom Tromey <tromey@adacore.com>
* ada-lang.c (ada_is_variant_part, ada_to_fixed_type_1):
Check ADA_TYPE_P.
(empty_record, ada_template_to_fixed_record_type_1)
(template_to_static_fixed_type)
(to_record_with_fixed_variant_part): Use INIT_NONE_SPECIFIC.
* cp-abi.c (value_rtti_type): Check HAVE_CPLUS_STRUCT.
* gdbtypes.h (INIT_NONE_SPECIFIC, ADA_TYPE_P): New
macros.
gdb/testsuite/ChangeLog
2019-04-19 Tom Tromey <tromey@adacore.com>
* gdb.ada/ptype_union.c: New file.
* gdb.ada/ptype_union.exp: New file.
This series is revisit of Siddhesh Poyarekar's patch from back in
2012. The last status on the patch is in the following gdb-patches
thread:
https://sourceware.org/ml/gdb-patches/2012-08/msg00562.html
It appears that Tom approved the patch, but Jan had some issues
with a compiler error that made the test fail on -m32 test runs.
He wrote up a hand-tweaked .S file to deal with it. Siddesh said
he would update tests. Then nothing.
Siddesh and Jan have both moved on since.
The patch originally required a large precursor patch to work.
I have whittled this down to/rewritten the bare minimum, and this
first patch is the result, changing the type of TYPE_LENGTH
to ULONGEST from unsigned int.
The majority of the changes involve changing printf format
strings to use %s and pulongest instead of %d.
gdb/ChangeLog:
* ada-lang.c (ada_template_to_fixed_record_type_1): Use
%s/pulongest for TYPE_LENGTH instead of %d in format
strings.
* ada-typerint.c (ada_print_type): Likewise.
* amd64-windows-tdep.c (amd64_windows_store_arg_in_reg): Likewise.
* compile/compile-c-support.c (generate_register_struct): Likewise.
* gdbtypes.c (recursive_dump_type): Likewise.
* gdbtypes.h (struct type) <length>: Change type to ULONGEST.
* m2-typeprint.c (m2_array): Use %s/pulongest for TYPE_LENGTH
instead of %d in format strings.
* riscv-tdep.c (riscv_type_alignment): Cast second argument
to std::min to ULONGEST.
* symmisc.c (print_symbol): Use %s/pulongest for TYPE_LENGTH
instead of %d in format strings.
* tracepoint.c (info_scope_command): Likewise.
* typeprint.c (print_offset_data::update)
(print_offset_data::finish): Likewise.
* xtensa-tdep.c (xtensa_store_return_value)
(xtensa_push_dummy_call): Likewise.
lookup_struct_elt is a new function which returns a tuple of
information about a component of a structure or union. The returned
tuple contains a pointer to the struct field object for the component
as well as a bit offset of that field within the structure. If the
field names a field in an anonymous substructure, the offset is the
"global" offset relative to the original structure type. If noerr is
set, then the returned tuple will set the field pointer to NULL to
indicate a missing component rather than throwing an error.
lookup_struct_elt_type is now reimplemented in terms of this new
function. It simply returns the type of the returned field.
gdb/ChangeLog:
* gdbtypes.c (lookup_struct_elt): New function.
(lookup_struct_elt_type): Reimplement via lookup_struct_elt.
* gdbtypes.h (struct struct_elt): New type.
(lookup_struct_elt): New prototype.
This commit applies all changes made after running the gdb/copyright.py
script.
Note that one file was flagged by the script, due to an invalid
copyright header
(gdb/unittests/basic_string_view/element_access/char/empty.cc).
As the file was copied from GCC's libstdc++-v3 testsuite, this commit
leaves this file untouched for the time being; a patch to fix the header
was sent to gcc-patches first.
gdb/ChangeLog:
Update copyright year range in all GDB files.
badness_vector is currently an open coded vector. This reimplements
it as a std::vector.
This fixes a few leaks as well:
- find_oload_champ is leaking every badness vector calculated bar the
one returned.
- bv->rank is always leaked, since callers of rank_function only
xfree the badness_vector pointer, not bv->rank.
gdb/ChangeLog:
2018-11-21 Pedro Alves <palves@redhat.com>
* gdbtypes.c (compare_badness): Change type of parameters to const
reference. Adjust to badness_vector being a std::vector now.
(rank_function): Adjust to badness_vector being a std::vector now.
* gdbtypes.h (badness_vector): Now a typedef to std::vector.
(LENGTH_MATCH): Delete.
(compare_badness): Change type of parameters to const reference.
(rank_function): Return a badness_vector by value now.
(find_overload_match): Adjust to badness_vector being a
std::vector now. Remove cleanups.
(find_oload_champ_namespace): 'oload_champ_bv' parameter now a
badness_vector pointer.
(find_oload_champ_namespace_loop): 'oload_champ_bv' parameter now
a badness_vector pointer. Adjust to badness_vector being a
std::vector now. Remove cleanups.
(find_oload_champ): 'oload_champ_bv' parameter now
a badness_vector pointer. Adjust to badness_vector being a
std::vector now. Remove cleanups.
This replaces more pointer+length with gdb::array_view. This time,
around invoke_xmethod, and then propagating the fallout around, which
inevitably leaks to the overload resolution code.
There are several places in the code that want to grab a slice of an
array, by advancing the array pointer, and decreasing the length
pointer. This patch introduces a pair of new
gdb::array_view::slice(...) methods to make that convenient and clear.
Unit test included.
gdb/ChangeLog:
2018-11-21 Pedro Alves <palves@redhat.com>
* common/array-view.h (array_view::splice(size_type, size_t)): New.
(array_view::splice(size_type)): New.
* eval.c (eval_call, evaluate_funcall): Adjust to use array_view.
* extension.c (xmethod_worker::get_arg_types): Adjust to return an
std::vector.
(xmethod_worker::get_result_type): Adjust to use gdb::array_view.
* extension.h: Include "common/array-view.h".
(xmethod_worker::invoke): Adjust to use gdb::array_view.
(xmethod_worker::get_arg_types): Adjust to return an std::vector.
(xmethod_worker::get_result_type): Adjust to use gdb::array_view.
(xmethod_worker::do_get_arg_types): Adjust to use std::vector.
(xmethod_worker::do_get_result_type): Adjust to use
gdb::array_view.
* gdbtypes.c (rank_function): Adjust to use gdb::array_view.
* gdbtypes.h: Include "common/array-view.h".
(rank_function): Adjust to use gdb::array_view.
* python/py-xmethods.c (python_xmethod_worker::invoke)
(python_xmethod_worker::do_get_arg_types)
(python_xmethod_worker::do_get_result_type)
(python_xmethod_worker::invoke): Adjust to new interfaces.
* valarith.c (value_user_defined_cpp_op, value_user_defined_op)
(value_x_binop, value_x_unop): Adjust to use gdb::array_view.
* valops.c (find_overload_match, find_oload_champ_namespace)
(find_oload_champ_namespace_loop, find_oload_champ): Adjust to use
gdb:array_view and the new xmethod_worker interfaces.
* value.c (result_type_of_xmethod, call_xmethod): Adjust to use
gdb::array_view.
* value.h (find_overload_match, result_type_of_xmethod)
(call_xmethod): Adjust to use gdb::array_view.
* unittests/array-view-selftests.c: Add slicing tests.
-fsanitize=undefined complains about using operator~ on various enum
types that are used with DEF_ENUM_FLAGS_TYPE. This patch fixes these
problems by explicitly setting the base type for these enums to
unsigned. It also adds a static assert to enum_flags to ensure that
future enums used this way have an unsigned underlying type.
gdb/ChangeLog
2018-10-03 Tom Tromey <tom@tromey.com>
* common/enum-flags.h (enum_flags::operator~): Add static assert.
* symfile-add-flags.h (enum symfile_add_flag): Use unsigned as
base type.
* objfile-flags.h (enum objfile_flag): Use unsigned as base type.
* gdbtypes.h (enum type_instance_flag_value): Use unsigned as base
type.
* c-lang.h (enum c_string_type_values): Use unsigned as base
type.
* btrace.h (enum btrace_thread_flag): Use unsigned as base type.
This patch started as an observation from valgrind that GDB appeared
to be loosing track of some memory associated with types. An example
valgrind stack would be:
24 bytes in 1 blocks are possibly lost in loss record 419 of 5,361
at 0x4C2EA1E: calloc (vg_replace_malloc.c:711)
by 0x623D26: xcalloc (common-utils.c:85)
by 0x623D65: xzalloc(unsigned long) (common-utils.c:95)
by 0x72A066: make_function_type(type*, type**) (gdbtypes.c:510)
by 0x72A098: lookup_function_type(type*) (gdbtypes.c:521)
by 0x73635D: gdbtypes_post_init(gdbarch*) (gdbtypes.c:5439)
by 0x727590: gdbarch_data(gdbarch*, gdbarch_data*) (gdbarch.c:5230)
by 0x735B99: builtin_type(gdbarch*) (gdbtypes.c:5313)
by 0x514D95: elf_rel_plt_read(minimal_symbol_reader&, objfile*, bfd_symbol**) (elfread.c:542)
by 0x51662F: elf_read_minimal_symbols(objfile*, int, elfinfo const*) (elfread.c:1121)
by 0x5168A5: elf_symfile_read(objfile*, enum_flags<symfile_add_flag>) (elfread.c:1207)
by 0x8520F5: read_symbols(objfile*, enum_flags<symfile_add_flag>) (symfile.c:794)
When we look in make_function_type we find a call to TYPE_ZALLOC
(inside the INIT_FUNC_SPECIFIC macro). It is this call to TYPE_ZALLOC
that is allocating memory with xcalloc, that is then getting lost.
The problem is tht calling TYPE_ALLOC or TYPE_ZALLOC currently
allocates memory from either the objfile obstack or by using malloc.
The problem with this is that types are allocated either on the
objfile obstack, or on the gdbarch obstack.
As a result, if we discard a type associated with an objfile then
auxiliary data allocated with TYPE_(Z)ALLOC will be correctly
discarded. But, if we were ever to discard a gdbarch then any
auxiliary type data would be leaked. Right now there are very few
places in GDB where a gdbarch is ever discarded, but it shouldn't hurt
to close down these bugs as we spot them.
This commit ensures that auxiliary type data is allocated from the
same obstack as the type itself, which should reduce leaked memory.
The one problem case that I found with this change was in eval.c,
where in one place we allocate a local type structure, and then used
TYPE_ZALLOC to allocate some space for the type. This local type is
neither object file owned, nor gdbarch owned, and so the updated
TYPE_ALLOC code is unable to find an objstack to allocate space on.
My proposed solution for this issue is that the space should be
allocated with a direct call to xzalloc. We could extend TYPE_ALLOC
to check for type->gdbarch being null, and then fall back to a direct
call to xzalloc, however, I think that making this rare case of a
local type require special handling is not a bad thing, this serves to
highlight that clearing up the memory will require special handling
too.
This special case of a local type is interesting as the types owner
field (contained within the main_type) is completely null. While
reflecting on this I looked at how types use the get_type_arch
function. It seems clear that, based on how this is used, it is never
intended that null will be returned from this function. This only
goes to reinforce, how locally alloctaed types, with no owner, are
both special, and need to be handled carefully. To help spot errors
earlier, I added an assert into get_type_arch that the returned arch
is not null.
Inside gdbarch.c I found a few other places where auxiliary type data
was being allocated directly on the heap rather than on the types
obstack. I have fixed these to call TYPE_ALLOC now.
Finally, it is worth noting that as we don't clean up our gdbarch
objects yet, then this will not make much of an impact on the amount
of memory reported as lost at program termination time. Memory
allocated for auxiliary type information is still not freed, however,
it is now on the correct obstack. If we do ever start freeing our
gdbarch structures then the associated type data will be cleaned up
correctly.
Tested on X86-64 GNU/Linux with no regressions.
gdb/ChangeLog:
* eval.c (fake_method::fake_method): Call xzalloc directly for a
type that is neither object file owned, nor gdbarch owned.
* gdbtypes.c (get_type_gdbarch): Add an assert that returned
gdbarch is non-NULL.
(alloc_type_instance): Allocate non-objfile owned types on the
gdbarch obstack.
(copy_type_recursive): Allocate TYPE_FIELDS and TYPE_RANGE_DATA
using TYPE_ALLOC to ensure memory is allocated on the correct
obstack.
* gdbtypes.h (TYPE_ALLOC): Allocate space on either the objfile
obstack, or the gdbarch obstack.
(TYPE_ZALLOC): Rewrite using TYPE_ALLOC.
Add int24 and uint24. These are used by the upcoming S12Z target, but will be
needed for any arch which features 24 bit registers.
* gdb/gdbtypes.h (struct builtin_type): New members builtin_int24
and builtin_uint24;
* gdb/gdbtypes.c: Initialize them.
* gdb/doc/gdb.texinfo (Predefined Target Types): Mention types int24 and uint24.
TYPE_TAG_NAME has been an occasional source of confusion and bugs. It
seems to me that it is only useful for C and C++ -- but even there,
not so much, because at least with DWARF there doesn't seem to be any
way to wind up with a type where the name and the tag name are both
non-NULL and different.
So, this patch removes TYPE_TAG_NAME entirely. This should save a
little memory, but more importantly, it simplifies this part of gdb.
A few minor test suite adjustments were needed. In some situations
the new code does not yield identical output to the old code.
gdb/ChangeLog
2018-06-01 Tom Tromey <tom@tromey.com>
* valops.c (enum_constant_from_type, value_namespace_elt)
(value_maybe_namespace_elt): Update.
* valarith.c (find_size_for_pointer_math): Update.
* target-descriptions.c (make_gdb_type): Update.
* symmisc.c (print_symbol): Update.
* stabsread.c (define_symbol, read_type)
(complain_about_struct_wipeout, add_undefined_type)
(cleanup_undefined_types_1): Update.
* rust-lang.c (rust_tuple_type_p, rust_slice_type_p)
(rust_range_type_p, val_print_struct, rust_print_struct_def)
(rust_internal_print_type, rust_composite_type)
(rust_evaluate_funcall, rust_evaluate_subexp)
(rust_inclusive_range_type_p): Update.
* python/py-type.c (typy_get_tag): Update.
* p-typeprint.c (pascal_type_print_base): Update.
* mdebugread.c (parse_symbol, parse_type): Update.
* m2-typeprint.c (m2_long_set, m2_record_fields, m2_enum):
Update.
* guile/scm-type.c (gdbscm_type_tag): Update.
* go-lang.c (sixg_string_p): Update.
* gnu-v3-abi.c (build_gdb_vtable_type, build_std_type_info_type):
Update.
* gdbtypes.h (struct main_type) <tag_name>: Remove.
(TYPE_TAG_NAME): Remove.
* gdbtypes.c (type_name_no_tag): Simplify.
(check_typedef, check_types_equal, recursive_dump_type)
(copy_type_recursive, arch_composite_type): Update.
* f-typeprint.c (f_type_print_base): Update. Print "Type" prefix
in summary mode when needed.
* eval.c (evaluate_funcall): Update.
* dwarf2read.c (fixup_go_packaging, read_structure_type)
(process_structure_scope, read_enumeration_type)
(read_namespace_type, read_module_type, determine_prefix): Update.
* cp-support.c (inspect_type): Update.
* coffread.c (process_coff_symbol, decode_base_type): Update.
* c-varobj.c (c_is_path_expr_parent): Update.
* c-typeprint.c (c_type_print_base_struct_union): Update.
(c_type_print_base_1): Update. Print struct/class/union/enum in
summary when using C language.
* ax-gdb.c (gen_struct_ref, gen_namespace_elt)
(gen_maybe_namespace_elt): Update.
* ada-lang.c (ada_type_name): Simplify.
(empty_record, ada_template_to_fixed_record_type_1)
(template_to_static_fixed_type)
(to_record_with_fixed_variant_part, ada_check_typedef): Update.
gdb/testsuite/ChangeLog
2018-06-01 Tom Tromey <tom@tromey.com>
* gdb.xml/tdesc-regs.exp (load_description): Update expected
results.
* gdb.dwarf2/method-ptr.exp: Set language to C++.
* gdb.dwarf2/member-ptr-forwardref.exp: Set language to C++.
* gdb.cp/typeid.exp (do_typeid_tests): Update type_re.
* gdb.base/maint.exp (maint_pass_if): Update.
This removes a VEC from type.c, by using std::vector.
While doing this I also took the opportunity to change
types_deeply_equal to return bool. This caught some weird code in
typy_richcompare, now fixed.
And, since I was changing types_deeply_equal, it seemed like a good
idea to also change types_equal, so this patch includes that as well.
Tested by the buildbot.
ChangeLog
2018-05-29 Tom Tromey <tom@tromey.com>
* python/py-type.c (typy_richcompare): Update.
* guile/scm-type.c (tyscm_equal_p_type_smob): Update.
* gdbtypes.h (types_deeply_equal): Return bool.
(types_equal): Likewise.
* gdbtypes.c (type_equality_entry_d): Remove typedef. Don't
declare VEC.
(check_types_equal): Change worklist to std::vector. Return
bool.
(struct type_equality_entry): Add constructor.
(compare_maybe_null_strings): Return bool.
(check_types_worklist): Return bool. Change worklist to
std::vector.
(types_deeply_equal): Use std::vector.
(types_equal): Return bool.
(compare_maybe_null_strings): Simplify.
This adds some basic type alignment support to gdb. It changes struct
type to store the alignment, and updates dwarf2read.c to handle
DW_AT_alignment. It also adds a new gdbarch method and updates
i386-tdep.c.
None of this new functionality is used anywhere yet, so tests will
wait until the next patch.
2018-04-30 Tom Tromey <tom@tromey.com>
* i386-tdep.c (i386_type_align): New function.
(i386_gdbarch_init): Update.
* gdbarch.sh (type_align): New method.
* gdbarch.c, gdbarch.h: Rebuild.
* arch-utils.h (default_type_align): Declare.
* arch-utils.c (default_type_align): New function.
* gdbtypes.h (TYPE_ALIGN_BITS): New define.
(struct type) <align_log2>: New field.
<instance_flags>: Now a bitfield.
(TYPE_RAW_ALIGN): New macro.
(type_align, type_raw_align, set_type_align): Declare.
* gdbtypes.c (type_align, type_raw_align, set_type_align): New
functions.
* dwarf2read.c (quirk_rust_enum): Set type alignment.
(get_alignment, maybe_set_alignment): New functions.
(read_structure_type, read_enumeration_type, read_array_type)
(read_set_type, read_tag_pointer_type, read_tag_reference_type)
(read_subrange_type, read_base_type): Set type alignment.
This adds some initial support for variant parts to gdbtypes.h. A
variant part is represented as a union. The union has a flag
indicating that it has a discriminant, and information about the
discriminant is attached using the dynamic property system.
2018-02-26 Tom Tromey <tom@tromey.com>
* value.h (value_union_variant): Declare.
* valops.c (value_union_variant): New function.
* gdbtypes.h (TYPE_FLAG_DISCRIMINATED_UNION): New macro.
(struct discriminant_info): New.
(enum dynamic_prop_node_kind) <DYN_PROP_DISCRIMINATED>: New
enumerator.
(struct main_type) <flag_discriminated_union>: New field.
Does anybody have an opinion about this? It would be nice to unbreak
the "default" build with clang (i.e. without passing special -Wno-error=
flags).
Here's a version rebased on today's master.
From 47d28075117fa2ddb93584ec50881e33777a85e5 Mon Sep 17 00:00:00 2001
From: Simon Marchi <simon.marchi@ericsson.com>
Date: Sat, 30 Dec 2017 22:48:18 -0500
Subject: [PATCH] dwarf: Make sect_offset 64-bits
Compiling with Clang 6 shows these errors:
/home/emaisin/src/binutils-gdb/gdb/dwarf2read.c:26610:43: error: result of comparison of constant 4294967296 with expression of type 'typename std::underlying_type<sect_offset>::type' (a
ka 'unsigned int') is always false [-Werror,-Wtautological-constant-out-of-range-compare]
if (to_underlying (per_cu.sect_off) >= (static_cast<uint64_t> (1) << 32))
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/home/emaisin/src/binutils-gdb/gdb/dwarf2read.c:26618:43: error: result of comparison of constant 4294967296 with expression of type 'typename std::underlying_type<sect_offset>::type' (a
ka 'unsigned int') is always false [-Werror,-Wtautological-constant-out-of-range-compare]
if (to_underlying (per_cu.sect_off) >= (static_cast<uint64_t> (1) << 32))
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The code in question checks if there is any offset exceeding 32 bits,
and therefore if we need to use the 64-bit DWARF format when writing the
.debug_names section. The type we use currently to represent section
offsets is an unsigned int (32-bits), which means a value of this type
will never exceed 32 bits, hence the errors above.
There are many signs that we want to support 64-bits DWARF (although I
haven't tested), such as:
- We correctly read initial length fields (read_initial_length)
- We take that into account when reading offsets (read_offset_1)
- The check_dwarf64_offsets function
However, I don't see how it can work if sect_offset is a 32-bits type.
Every time we record a section offset, we risk truncating the value.
And if a file uses the 64-bit DWARF format, it's most likely because
there are such offset values that overflow 32 bits.
Because of this, I think the way forward is to change sect_offset to be
a uint64_t. It will be able to represent any offset, regardless of the
bitness of the DWARF info.
This patch was regtested on the buildbot.
gdb/ChangeLog:
* gdbtypes.h (sect_offset): Change type to uint64_t.
(sect_offset_str): New function.
* dwarf2read.c (create_addrmap_from_aranges): Use
sect_offset_str.
(error_check_comp_unit_head): Likewise.
(create_debug_type_hash_table): Likewise.
(read_cutu_die_from_dwo): Likewise.
(init_cutu_and_read_dies): Likewise.
(init_cutu_and_read_dies_no_follow): Likewise.
(process_psymtab_comp_unit_reader): Likewise.
(partial_die_parent_scope): Likewise.
(peek_die_abbrev): Likewise.
(process_queue): Likewise.
(dwarf2_physname): Likewise.
(read_namespace_alias): Likewise.
(read_import_statement): Likewise.
(create_dwo_cu_reader): Likewise.
(create_cus_hash_table): Likewise.
(lookup_dwo_cutu): Likewise.
(inherit_abstract_dies): Likewise.
(read_func_scope): Likewise.
(read_call_site_scope): Likewise.
(dwarf2_add_member_fn): Likewise.
(read_common_block): Likewise.
(read_module_type): Likewise.
(read_typedef): Likewise.
(read_subrange_type): Likewise.
(load_partial_dies): Likewise.
(read_partial_die): Likewise.
(find_partial_die): Likewise.
(read_str_index): Likewise.
(dwarf2_string_attr): Likewise.
(build_error_marker_type): Likewise.
(lookup_die_type): Likewise.
(dump_die_shallow): Likewise.
(follow_die_ref): Likewise.
(dwarf2_fetch_die_loc_sect_off): Likewise.
(dwarf2_fetch_constant_bytes): Likewise.
(follow_die_sig): Likewise.
(get_signatured_type): Likewise.
(get_DW_AT_signature_type): Likewise.
(dwarf2_find_containing_comp_unit): Likewise.
(set_die_type): Likewise.
GCC PR83906 [1] is about a GCC/libstdc++ GDB/Python type printer
testcase failing randomly, as shown by running (in libstdc++'s
testsuite):
make check RUNTESTFLAGS=prettyprinters.exp=80276.cc
in a loop. Sometimes you get this:
FAIL: libstdc++-prettyprinters/80276.cc whatis p4
I.e., this:
type = std::unique_ptr<std::vector<std::unique_ptr<std::list<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >>[]>>[99]>
instead of this:
type = std::unique_ptr<std::vector<std::unique_ptr<std::list<std::string>[]>>[99]>
Jonathan Wakely tracked it on the printer side to this bit in
libstdc++'s type printer:
if self.type_obj == type_obj:
return strip_inline_namespaces(self.name)
This assumes the two types resolve to the same gdb.Type but some times
the comparison unexpectedly fails.
Running the testcase manually under Valgrind finds the problem in GDB:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~
==6118== Conditional jump or move depends on uninitialised value(s)
==6118== at 0x4C35CB0: bcmp (vg_replace_strmem.c:1100)
==6118== by 0x6F773A: check_types_equal(type*, type*, VEC_type_equality_entry_d**) (gdbtypes.c:3515)
==6118== by 0x6F7B00: check_types_worklist(VEC_type_equality_entry_d**, bcache*) (gdbtypes.c:3618)
==6118== by 0x6F7C03: types_deeply_equal(type*, type*) (gdbtypes.c:3655)
==6118== by 0x4D5B06: typy_richcompare(_object*, _object*, int) (py-type.c:1007)
==6118== by 0x63D7E6C: PyObject_RichCompare (object.c:961)
==6118== by 0x646EAEC: PyEval_EvalFrameEx (ceval.c:4960)
==6118== by 0x646DC08: PyEval_EvalFrameEx (ceval.c:4519)
==6118== by 0x646DC08: PyEval_EvalFrameEx (ceval.c:4519)
==6118== by 0x646DC08: PyEval_EvalFrameEx (ceval.c:4519)
==6118== by 0x646DC08: PyEval_EvalFrameEx (ceval.c:4519)
==6118== by 0x646DC08: PyEval_EvalFrameEx (ceval.c:4519)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~
That "bcmp" call is really a memcmp call in check_types_equal. The
problem is that gdb is memcmp'ing two objects that are equal in value:
(top-gdb) p *TYPE_RANGE_DATA (type1)
$1 = {low = {kind = PROP_CONST, data = {const_val = 0, baton = 0x0}}, high = {kind = PROP_CONST, data = {const_val = 15, baton = 0xf}}, flag_upper_bound_is_count = 0,
flag_bound_evaluated = 0}
(top-gdb) p *TYPE_RANGE_DATA (type2)
$2 = {low = {kind = PROP_CONST, data = {const_val = 0, baton = 0x0}}, high = {kind = PROP_CONST, data = {const_val = 15, baton = 0xf}}, flag_upper_bound_is_count = 0,
flag_bound_evaluated = 0}
but differ in padding. Notice the 4-byte hole:
(top-gdb) ptype /o range_bounds
/* offset | size */ type = struct range_bounds {
/* 0 | 16 */ struct dynamic_prop {
/* 0 | 4 */ dynamic_prop_kind kind;
/* XXX 4-byte hole */
/* 8 | 8 */ union dynamic_prop_data {
/* 8 */ LONGEST const_val;
/* 8 */ void *baton;
/* total size (bytes): 8 */
} data;
which is filled with garbage:
(top-gdb) x /40bx TYPE_RANGE_DATA (type1)
0x2fa7ea0: 0x01 0x00 0x00 0x00 0x43 0x01 0x00 0x00
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
0x2fa7ea8: 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00
0x2fa7eb0: 0x01 0x00 0x00 0x00 0xfe 0x7f 0x00 0x00
0x2fa7eb8: 0x0f 0x00 0x00 0x00 0x00 0x00 0x00 0x00
0x2fa7ec0: 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00
(top-gdb) x /40bx TYPE_RANGE_DATA (type2)
0x20379b0: 0x01 0x00 0x00 0x00 0xfe 0x7f 0x00 0x00
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
0x20379b8: 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00
0x20379c0: 0x01 0x00 0x00 0x00 0xfe 0x7f 0x00 0x00
0x20379c8: 0x0f 0x00 0x00 0x00 0x00 0x00 0x00 0x00
0x20379d0: 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00
(top-gdb) p memcmp (TYPE_RANGE_DATA (type1), TYPE_RANGE_DATA (type2), sizeof (*TYPE_RANGE_DATA (type1)))
$3 = -187
In some cases objects of type range_bounds are memset when allocated,
but then their dynamic_prop low/high fields are copied over from some
template dynamic_prop object that wasn't memset. E.g.,
create_static_range_type's low/high locals are left with garbage in
the padding, and then that padding is copied over to the range_bounds
object's low/high fields.
At first, I considered making sure to always memset range_bounds
objects, thinking that maybe type objects are being put in some bcache
instance somewhere. But then I hacked bcache/bcache_full to poison
non-pod types, and made dynamic_prop a non-pod, and GDB still
compiled.
So given that, it seems safest to not assume padding will always be
memset, and instead treat them as regular value types, implementing
(in)equality operators and using those instead of memcmp.
This fixes the random FAILs in GCC's testcase.
[1] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83906
gdb/ChangeLog:
2018-01-24 Pedro Alves <palves@redhat.com>
GCC PR libstdc++/83906
* gdbtypes.c (operator==(const dynamic_prop &,
const dynamic_prop &)): New.
(operator==(const range_bounds &, const range_bounds &)): New.
(check_types_equal): Use them instead of memcmp.
* gdbtypes.h (operator==(const dynamic_prop &,
const dynamic_prop &)): Declare.
(operator!=(const dynamic_prop &, const dynamic_prop &)): Declare.
(operator==(const range_bounds &, const range_bounds &)): Declare.
(operator!=(const range_bounds &, const range_bounds &)): Declare.
The objfile argument to add_dyn_prop is redundant, so this patch
removes it.
2018-01-17 Tom Tromey <tom@tromey.com>
* gdbtypes.h (add_dyn_prop): Remove objfile parameter.
* gdbtypes.c (add_dyn_prop): Remove objfile parameter.
(create_array_type_with_stride): Update.
* dwarf2read.c (set_die_type): Update.
This patch adds support for DW_AT_byte_stride, using Ada as one
example of where this would be useful. However, the implementation
is language-agnostic.
Consider the following Ada code:
procedure Nested (L, U : Integer) is
subtype Small_Type is Integer range L .. U;
type Record_Type (I : Small_Type := L) is record
S : String (1 .. I);
end record;
type Array_Type is array (Integer range <>) of Record_Type;
A1 : Array_Type :=
(1 => (I => 0, S => <>),
2 => (I => 1, S => "A"),
3 => (I => 2, S => "AB"));
procedure Discard (R : Record_Type) is
begin
null;
end Discard;
begin
Discard (A1 (1)); -- STOP
end;
It defines an array A1 of Record_Type, which is a variant record
type whose maximum size actually depends on the value of the
parameters passed when calling Nested. As a result, the stride
of the array A1 cannot be known statically, which leads the compiler
to generate a dynamic DW_AT_byte_stride attribute for our type.
Here is what the debugging info looks like with GNAT:
.uleb128 0x10 # (DIE (0x14e) DW_TAG_array_type)
.long .LASF17 # DW_AT_name: "foo__nested__T18b"
.long 0x141 # DW_AT_byte_stride
.long 0xdc # DW_AT_type
.uleb128 0x11 # (DIE (0x15f) DW_TAG_subrange_type)
.long 0x166 # DW_AT_type
.byte 0x3 # DW_AT_upper_bound
.byte 0 # end of children of DIE 0x14e
There DW_AT_byte_stride is a reference to a local (internal)
variable:
.uleb128 0x9 # (DIE (0x141) DW_TAG_variable)
.long .LASF6 # DW_AT_name: "foo__nested__T18b___PAD___XVZ"
This patch enhances GDB to handle this dynamic byte stride attribute
by first adding a new dynamic_prop_node_kind (DYN_PROP_BYTE_STRIDE)
to store the array dynamic stride info (when dynamic). It then enhances
the dynamic type resolver to handle this dynamic property.
Before applying this patch, trying to print the value of some of
A1's elements after having stopped at the "STOP" comment does not
work. For instance:
(gdb) p a1(2)
Cannot access memory at address 0x80000268dec0
With this patch applied, GDB now prints the value of all 3 elements
correctly:
(gdb) print A1(1)
$1 = (i => 0, s => "")
(gdb) print A1(2)
$2 = (i => 1, s => "A")
(gdb) print A1(3)
$3 = (i => 2, s => "AB")
gdb/ChangeLog:
* gdbtypes.h (enum dynamic_prop_node_kind) <DYN_PROP_BYTE_STRIDE>:
New enum value.
(create_array_type_with_stride): Add byte_stride_prop parameter.
* gdbtypes.c (create_array_type_with_stride) <byte_stride_prop>:
New parameter. Update all callers in this file.
(array_type_has_dynamic_stride): New function.
(is_dynamic_type_internal, resolve_dynamic_array): Add handling
of arrays with dynamic byte strides.
* dwarf2read.c (read_array_type): Add support for dynamic
DW_AT_byte_stride attributes.
gdb/testsuite/ChangeLog:
* gdb.ada/dyn_stride: New testcase.
Tested on x86_64-linux.
GDB currently does not track types defined in classes. Consider:
class A
{
public:
class B
{
public:
class C { };
};
};
(gdb) ptype A
type = class A {
<no data fields>
}
This patch changes this behavior so that GDB records these nested types
and displays them to the user when he has set the (new) "print type"
option "nested-type-limit."
Example:
(gdb) set print type nested-type-limit 1
(gdb) ptype A
type = class A {
<no data fields>
class A::B {
<no data fields>
};
}
(gdb) set print type nested-type-limit 2
type = class A {
<no data fields>
class A::B {
<no data fields>
class A::B::C {
<no data fields>
};
};
}
By default, the code maintains the status quo, that is, it will not print
any nested type definitions at all.
Testing is carried out via cp_ptype_class which required quite a bit of
modification to permit recursive calling (for the nested types). This
was most easily facilitated by turning the ptype command output into a
queue. Upshot: the test suite now has stack and queue data structures that
may be used by test writers.
gdb/ChangeLog
* NEWS (New commands): Mention set/show print type nested-type-limit.
* c-typeprint.c (c_type_print_base): Print out nested types.
* dwarf2read.c (struct typedef_field_list): Rename to ...
(struct decl_field_list): ... this. Change all uses.
(struct field_info) <nested_types_list, nested_types_list_count>:
New fields.
(add_partial_symbol): Look for nested type definitions in C++, too.
(dwarf2_add_typedef): Rename to ...
(dwarf2_add_type_defn): ... this.
(type_can_define_types): New function.
Update assertion to use type_can_define_types.
Permit NULL for a field's name.
(process_structure_scope): Handle child DIEs of types that can
define types.
Copy the list of nested types into the type struct.
* gdbtypes.h (struct typedef_field): Rename to ...
(struct decl_field): ... this. Change all uses.
[is_protected, is_private]: New fields.
(struct cplus_struct_type) <nested_types, nested_types_count>: New
fields.
(TYPE_NESTED_TYPES_ARRAY, TYPE_NESTED_TYPES_FIELD)
(TYPE_NESTED_TYPES_FIELD_NAME, TYPE_NESTED_TYPES_FIELD_TYPE)
(TYPE_NESTED_TYPES_COUNT, TYPE_NESTED_TYPES_FIELD_PROTECTED)
(TYPE_NESTED_TYPES_FIELD_PRIVATE): New macros.
* typeprint.c (type_print_raw_options, default_ptype_flags): Add
default value for print_nested_type_limit.
(print_nested_type_limit): New static variable.
(set_print_type_nested_types, show_print_type_nested_types): New
functions.
(_initialize_typeprint): Register new commands for set/show
`print-nested-type-limit'.
* typeprint.h (struct type_print_options) [print_nested_type_limit]:
New field.
gdb/testsuite/ChangeLog
* gdb.cp/nested-types.cc: New file.
* gdb.cp/nested-types.exp: New file.
* lib/cp-support.exp: Load data-structures.exp library.
(debug_cp_test_ptype_class): New global.
(cp_ptype_class_verbose, next_line): New procedures.
(cp_test_ptype_class): Add and document new parameter `recursive_qid'.
Add and document new return value.
Switch the list of lines to a queue.
Add support for new `type' key for nested type definitions.
Add debugging/troubleshooting messages.
* lib/data-structures.exp: New file.
gdb/doc/ChangeLog
* gdb.texinfo (Symbols): Document "set print type nested-type-limit"
and "show print type nested-type-limit".
