FreeChainXenon/binutils-gdb
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

gdb: avoid resolving dynamic properties for non-allocated arrays

Browse source 
In PR gdb/27059 an issue was discovered where GDB would sometimes
trigger undefined behaviour in the form of signed integer overflow.
The problem here is that GDB was reading random garbage from the
inferior memory space, assuming this data was valid, and performing
arithmetic on it.

This bug raises an interesting general problem with GDB's DWARF
expression evaluator, which is this:

We currently assume that the DWARF expressions being evaluated are
well formed, and well behaving.  As an example, this is the expression
that the bug was running into problems on, this was used as the
expression for a DW_AT_byte_stride of a DW_TAG_subrange_type:

	DW_OP_push_object_address;
	DW_OP_plus_uconst: 88;
	DW_OP_deref;
	DW_OP_push_object_address;
	DW_OP_plus_uconst: 32;
	DW_OP_deref;
	DW_OP_mul

Two values are read from the inferior and multiplied together.  GDB
should not assume that any value read from the inferior is in any way
sane, as such the implementation of DW_OP_mul should be guarding
against overflow and doing something semi-sane here.

However, it turns out that the original bug PR gdb/27059, is hitting a
more specific case, which doesn't require changes to the DWARF
expression evaluator, so I'm going to leave the above issue for
another day.

In the test mentioned in the bug GDB is actually trying to resolve the
dynamic type of a Fortran array that is NOT allocated.  A
non-allocated Fortran array is one that does not have any data
allocated for it yet, and even the upper and lower bounds of the array
are not yet known.

It turns out that, at least for gfortran compiled code, the data
fields that describe the byte-stride are not initialised until the
array is allocated.

This leads me to the following conclusion: GDB should not try to
resolve the bounds, or stride information for an array that is not
allocated (or not associated, a similar, but slightly different
Fortran feature).  Instead, each of these properties should be set to
undefined if the array is not allocated (or associated).

That is what this commit does.  There's a new flag that is passed
around during the dynamic array resolution.  When this flag is true
the dynamic properties are resolved using the DWARF expressions as
they currently are, but when this flag is false the expressions are
not evaluated, and instead the properties are set to undefined.

gdb/ChangeLog:

	PR gdb/27059
	* eval.c (evaluate_subexp_for_sizeof): Handle not allocated and
	not associated arrays.
	* f-lang.c (fortran_adjust_dynamic_array_base_address_hack): Don't
	adjust arrays that are not allocated/associated.
	* gdbtypes.c (resolve_dynamic_range): Update header comment.  Add
	new parameter which is used to sometimes set dynamic properties to
	undefined.
	(resolve_dynamic_array_or_string): Update header comment.  Add new
	parameter which is used to guard evaluating dynamic properties.
	Resolve allocated/associated properties first.

gdb/testsuite/ChangeLog:

	PR gdb/27059
	* gdb.dwarf2/dyn-type-unallocated.c: New file.
	* gdb.dwarf2/dyn-type-unallocated.exp: New file.
This commit is contained in:
Andrew Burgess 2020-12-18 11:59:54 +00:00
parent 5ba3b20ec2
commit b7874836c3
7 changed files with 240 additions and 24 deletions
Download patch file Download diff file Expand all files Collapse all files

69
gdb/gdbtypes.c
View file

@ -2177,11 +2177,20 @@ static struct type *resolve_dynamic_type_internal
/* Given a dynamic range type (dyn_range_type) and a stack of
struct property_addr_info elements, return a static version
of that type. */
of that type.
When RESOLVE_P is true then the returned static range is created by
actually evaluating any dynamic properties within the range type, while
when RESOLVE_P is false the returned static range has all of the bounds
and stride information set to undefined. The RESOLVE_P set to false
case will be used when evaluating a dynamic array that is not
allocated, or not associated, i.e. the bounds information might not be
initialized yet. */
static struct type *
resolve_dynamic_range (struct type *dyn_range_type,
struct property_addr_info *addr_stack)
struct property_addr_info *addr_stack,
bool resolve_p = true)
{
CORE_ADDR value;
struct type *static_range_type, *static_target_type;
@ -2190,13 +2199,13 @@ resolve_dynamic_range (struct type *dyn_range_type,
gdb_assert (dyn_range_type->code () == TYPE_CODE_RANGE);
const struct dynamic_prop *prop = &dyn_range_type->bounds ()->low;
if (dwarf2_evaluate_property (prop, NULL, addr_stack, &value))
if (resolve_p && dwarf2_evaluate_property (prop, NULL, addr_stack, &value))
low_bound.set_const_val (value);
else
low_bound.set_undefined ();
prop = &dyn_range_type->bounds ()->high;
if (dwarf2_evaluate_property (prop, NULL, addr_stack, &value))
if (resolve_p && dwarf2_evaluate_property (prop, NULL, addr_stack, &value))
{
high_bound.set_const_val (value);
@ -2209,7 +2218,7 @@ resolve_dynamic_range (struct type *dyn_range_type,
bool byte_stride_p = dyn_range_type->bounds ()->flag_is_byte_stride;
prop = &dyn_range_type->bounds ()->stride;
if (dwarf2_evaluate_property (prop, NULL, addr_stack, &value))
if (resolve_p && dwarf2_evaluate_property (prop, NULL, addr_stack, &value))
{
stride.set_const_val (value);
@ -2242,11 +2251,16 @@ resolve_dynamic_range (struct type *dyn_range_type,
/* Resolves dynamic bound values of an array or string type TYPE to static
ones. ADDR_STACK is a stack of struct property_addr_info to be used if
needed during the dynamic resolution. */
needed during the dynamic resolution.
When RESOLVE_P is true then the dynamic properties of TYPE are
evaluated, otherwise the dynamic properties of TYPE are not evaluated,
instead we assume the array is not allocated/associated yet. */
static struct type *
resolve_dynamic_array_or_string (struct type *type,
struct property_addr_info *addr_stack)
struct property_addr_info *addr_stack,
bool resolve_p = true)
{
CORE_ADDR value;
struct type *elt_type;
@ -2262,29 +2276,44 @@ resolve_dynamic_array_or_string (struct type *type,
type = copy_type (type);
elt_type = type;
range_type = check_typedef (elt_type->index_type ());
range_type = resolve_dynamic_range (range_type, addr_stack);
/* Resolve allocated/associated here before creating a new array type, which
will update the length of the array accordingly. */
/* Resolve the allocated and associated properties before doing anything
else. If an array is not allocated or not associated then (at least
for Fortran) there is no guarantee that the data to define the upper
bound, lower bound, or stride will be correct. If RESOLVE_P is
already false at this point then this is not the first dimension of
the array and a more outer dimension has already marked this array as
not allocated/associated, as such we just ignore this property. This
is fine as GDB only checks the allocated/associated on the outer most
dimension of the array. */
prop = TYPE_ALLOCATED_PROP (type);
if (prop != NULL && dwarf2_evaluate_property (prop, NULL, addr_stack, &value))
prop->set_const_val (value);
if (prop != NULL && resolve_p
&& dwarf2_evaluate_property (prop, NULL, addr_stack, &value))
{
prop->set_const_val (value);
if (value == 0)
resolve_p = false;
}
prop = TYPE_ASSOCIATED_PROP (type);
if (prop != NULL && dwarf2_evaluate_property (prop, NULL, addr_stack, &value))
prop->set_const_val (value);
if (prop != NULL && resolve_p
&& dwarf2_evaluate_property (prop, NULL, addr_stack, &value))
{
prop->set_const_val (value);
if (value == 0)
resolve_p = false;
}
ary_dim = check_typedef (TYPE_TARGET_TYPE (elt_type));
range_type = check_typedef (type->index_type ());
range_type = resolve_dynamic_range (range_type, addr_stack, resolve_p);
ary_dim = check_typedef (TYPE_TARGET_TYPE (type));
if (ary_dim != NULL && ary_dim->code () == TYPE_CODE_ARRAY)
elt_type = resolve_dynamic_array_or_string (ary_dim, addr_stack);
elt_type = resolve_dynamic_array_or_string (ary_dim, addr_stack, resolve_p);
else
elt_type = TYPE_TARGET_TYPE (type);
prop = type->dyn_prop (DYN_PROP_BYTE_STRIDE);
if (prop != NULL)
if (prop != NULL && resolve_p)
{
if (dwarf2_evaluate_property (prop, NULL, addr_stack, &value))
{