diff mbox series

PR fortran/102332 - ICE in select_type_set_tmp, at fortran/match.c:6366

Message ID trinity-6d30d506-bd21-4123-a5fc-9457c7f2cca6-1640643461318@3c-app-gmx-bs23
State New
Headers show
Series PR fortran/102332 - ICE in select_type_set_tmp, at fortran/match.c:6366 | expand

Commit Message

Harald Anlauf Dec. 27, 2021, 10:17 p.m. UTC 
Dear all,

there are a couple of NULL pointer dereferences leading to improper
error recovery when trying to handle Gerhard's testcases involving
SELECT TYPE and invalid uses of CLASS variables.

The fixes look pretty obvious to me, but I'm submitting here to
check if there is more that should be done here.

(I was surprised to see that there are several different places
involved by rather simple variations in the basic test case.)

Regtested on x86_64-pc-linux-gnu.  OK for mainline?

Thanks,
Harald

Comments

Paul Richard Thomas Dec. 28, 2021, 11:56 a.m. UTC | #1 
Hi Harald,

This looks good to me. OK for mainline and, dare I suggest, 11-branch?

From a quick run through resolve.c, there are many places where the extra
checks that you introduced in the patch have been implemented. This makes
me wonder whether a function or macro might not make the relevant code more
concise.

Thanks for the patch

Paul


On Mon, 27 Dec 2021 at 22:17, Harald Anlauf via Fortran <fortran@gcc.gnu.org>
wrote:

> Dear all,
>
> there are a couple of NULL pointer dereferences leading to improper
> error recovery when trying to handle Gerhard's testcases involving
> SELECT TYPE and invalid uses of CLASS variables.
>
> The fixes look pretty obvious to me, but I'm submitting here to
> check if there is more that should be done here.
>
> (I was surprised to see that there are several different places
> involved by rather simple variations in the basic test case.)
>
> Regtested on x86_64-pc-linux-gnu.  OK for mainline?
>
> Thanks,
> Harald
>
>
Harald Anlauf Dec. 28, 2021, 9:08 p.m. UTC | #2 
Hi Paul,

Am 28.12.21 um 12:56 schrieb Paul Richard Thomas via Fortran:
> Hi Harald,
>
> This looks good to me. OK for mainline and, dare I suggest, 11-branch?
>
>  From a quick run through resolve.c, there are many places where the extra
> checks that you introduced in the patch have been implemented. This makes
> me wonder whether a function or macro might not make the relevant code more
> concise.

I had thought about this in the past, too.  Suitably chosen macros
could help to make checking not only more concise, but also more
robust and (hopefully) readable at the same time.

What do you think about e.g.


diff --git a/gcc/fortran/gfortran.h b/gcc/fortran/gfortran.h
index e5d2dd7971e..f3d22b46a75 100644
--- a/gcc/fortran/gfortran.h
+++ b/gcc/fortran/gfortran.h
@@ -3885,6 +3885,8 @@ bool gfc_is_finalizable (gfc_symbol *, gfc_expr **);
          && CLASS_DATA (sym) \
          && CLASS_DATA (sym)->attr.dimension \
          && !CLASS_DATA (sym)->attr.class_pointer)
+#define IS_CLASS_OBJ(sym) \
+       (sym->ts.type == BT_CLASS && sym->attr.class_ok)

  /* frontend-passes.c */


to be used to ensure that we are dealing with a CLASS object where
attributes should already have been set up?  Or use a better name?
(IS_CLASS_OBJECT?)

Thanks,
Harald

> Thanks for the patch
>
> Paul
>
>
> On Mon, 27 Dec 2021 at 22:17, Harald Anlauf via Fortran <fortran@gcc.gnu.org>
> wrote:
>
>> Dear all,
>>
>> there are a couple of NULL pointer dereferences leading to improper
>> error recovery when trying to handle Gerhard's testcases involving
>> SELECT TYPE and invalid uses of CLASS variables.
>>
>> The fixes look pretty obvious to me, but I'm submitting here to
>> check if there is more that should be done here.
>>
>> (I was surprised to see that there are several different places
>> involved by rather simple variations in the basic test case.)
>>
>> Regtested on x86_64-pc-linux-gnu.  OK for mainline?
>>
>> Thanks,
>> Harald
>>
>>
>
Paul Richard Thomas Dec. 29, 2021, 11:45 a.m. UTC | #3 
Hi Harald,

That is the sort of thing that I had in mind. Is it worth adding the check
for CLASS_DATA? I cannot remember if that is made redundant by the test of
the class_ok attribute.

Cheers

Paul


On Tue, 28 Dec 2021 at 21:08, Harald Anlauf <anlauf@gmx.de> wrote:

> Hi Paul,
>
> Am 28.12.21 um 12:56 schrieb Paul Richard Thomas via Fortran:
> > Hi Harald,
> >
> > This looks good to me. OK for mainline and, dare I suggest, 11-branch?
> >
> >  From a quick run through resolve.c, there are many places where the
> extra
> > checks that you introduced in the patch have been implemented. This makes
> > me wonder whether a function or macro might not make the relevant code
> more
> > concise.
>
> I had thought about this in the past, too.  Suitably chosen macros
> could help to make checking not only more concise, but also more
> robust and (hopefully) readable at the same time.
>
> What do you think about e.g.
>
>
> diff --git a/gcc/fortran/gfortran.h b/gcc/fortran/gfortran.h
> index e5d2dd7971e..f3d22b46a75 100644
> --- a/gcc/fortran/gfortran.h
> +++ b/gcc/fortran/gfortran.h
> @@ -3885,6 +3885,8 @@ bool gfc_is_finalizable (gfc_symbol *, gfc_expr **);
>           && CLASS_DATA (sym) \
>           && CLASS_DATA (sym)->attr.dimension \
>           && !CLASS_DATA (sym)->attr.class_pointer)
> +#define IS_CLASS_OBJ(sym) \
> +       (sym->ts.type == BT_CLASS && sym->attr.class_ok)
>
>   /* frontend-passes.c */
>
>
> to be used to ensure that we are dealing with a CLASS object where
> attributes should already have been set up?  Or use a better name?
> (IS_CLASS_OBJECT?)
>
> Thanks,
> Harald
>
> > Thanks for the patch
> >
> > Paul
> >
> >
> > On Mon, 27 Dec 2021 at 22:17, Harald Anlauf via Fortran <
> fortran@gcc.gnu.org>
> > wrote:
> >
> >> Dear all,
> >>
> >> there are a couple of NULL pointer dereferences leading to improper
> >> error recovery when trying to handle Gerhard's testcases involving
> >> SELECT TYPE and invalid uses of CLASS variables.
> >>
> >> The fixes look pretty obvious to me, but I'm submitting here to
> >> check if there is more that should be done here.
> >>
> >> (I was surprised to see that there are several different places
> >> involved by rather simple variations in the basic test case.)
> >>
> >> Regtested on x86_64-pc-linux-gnu.  OK for mainline?
> >>
> >> Thanks,
> >> Harald
> >>
> >>
> >
>
>
Harald Anlauf Dec. 29, 2021, 5:25 p.m. UTC | #4 
Hi Paul,

Am 29.12.21 um 12:45 schrieb Paul Richard Thomas via Fortran:
> Hi Harald,
>
> That is the sort of thing that I had in mind. Is it worth adding the check
> for CLASS_DATA? I cannot remember if that is made redundant by the test of
> the class_ok attribute.

the macro CLASS_DATA appears to have been introduced by Janus
as a shortcut to the data component.  Accesses to the data
component still may require a protecting prior check when we
are confronted with random bad user code, as I sometimes see
sym->attr.class_ok == 1 but still sym->ts.u.derived == NULL.

Shall we open a PR for tracking issues with (the lack of)
canonical checking when CLASS is involved?  Where we could
collect ideas about macros that make the code more readable
and robust?  (I believe some PRs involving "associate" need
similar treatment in several places.)

Of course that would be something for post 12-release.

Cheers,
Harald

> Cheers
>
> Paul
>
>
> On Tue, 28 Dec 2021 at 21:08, Harald Anlauf <anlauf@gmx.de> wrote:
>
>> Hi Paul,
>>
>> Am 28.12.21 um 12:56 schrieb Paul Richard Thomas via Fortran:
>>> Hi Harald,
>>>
>>> This looks good to me. OK for mainline and, dare I suggest, 11-branch?
>>>
>>>   From a quick run through resolve.c, there are many places where the
>> extra
>>> checks that you introduced in the patch have been implemented. This makes
>>> me wonder whether a function or macro might not make the relevant code
>> more
>>> concise.
>>
>> I had thought about this in the past, too.  Suitably chosen macros
>> could help to make checking not only more concise, but also more
>> robust and (hopefully) readable at the same time.
>>
>> What do you think about e.g.
>>
>>
>> diff --git a/gcc/fortran/gfortran.h b/gcc/fortran/gfortran.h
>> index e5d2dd7971e..f3d22b46a75 100644
>> --- a/gcc/fortran/gfortran.h
>> +++ b/gcc/fortran/gfortran.h
>> @@ -3885,6 +3885,8 @@ bool gfc_is_finalizable (gfc_symbol *, gfc_expr **);
>>            && CLASS_DATA (sym) \
>>            && CLASS_DATA (sym)->attr.dimension \
>>            && !CLASS_DATA (sym)->attr.class_pointer)
>> +#define IS_CLASS_OBJ(sym) \
>> +       (sym->ts.type == BT_CLASS && sym->attr.class_ok)
>>
>>    /* frontend-passes.c */
>>
>>
>> to be used to ensure that we are dealing with a CLASS object where
>> attributes should already have been set up?  Or use a better name?
>> (IS_CLASS_OBJECT?)
>>
>> Thanks,
>> Harald
>>
>>> Thanks for the patch
>>>
>>> Paul
>>>
>>>
>>> On Mon, 27 Dec 2021 at 22:17, Harald Anlauf via Fortran <
>> fortran@gcc.gnu.org>
>>> wrote:
>>>
>>>> Dear all,
>>>>
>>>> there are a couple of NULL pointer dereferences leading to improper
>>>> error recovery when trying to handle Gerhard's testcases involving
>>>> SELECT TYPE and invalid uses of CLASS variables.
>>>>
>>>> The fixes look pretty obvious to me, but I'm submitting here to
>>>> check if there is more that should be done here.
>>>>
>>>> (I was surprised to see that there are several different places
>>>> involved by rather simple variations in the basic test case.)
>>>>
>>>> Regtested on x86_64-pc-linux-gnu.  OK for mainline?
>>>>
>>>> Thanks,
>>>> Harald
>>>>
>>>>
>>>
>>
>>
>
diff mbox series

Patch

From 4cda248202ea741bea1dd1ca4531aa15f423801b Mon Sep 17 00:00:00 2001
From: Harald Anlauf <anlauf@gmx.de>
Date: Mon, 27 Dec 2021 23:06:18 +0100
Subject: [PATCH] Fortran: avoid several NULL pointer dereferences during error
 recovery

gcc/fortran/ChangeLog:

	PR fortran/102332
	* expr.c (gfc_get_variable_expr): Avoid NULL pointer dereferences
	during handling of errors with invalid uses of CLASS variables.
	* match.c (select_type_set_tmp): Likewise.
	* primary.c (gfc_match_varspec): Likewise.
	* resolve.c (resolve_variable): Likewise.
	(resolve_select_type): Likewise.

gcc/testsuite/ChangeLog:

	PR fortran/102332
	* gfortran.dg/pr102332.f90: New test.
---
 gcc/fortran/expr.c                     |  3 +-
 gcc/fortran/match.c                    |  3 +-
 gcc/fortran/primary.c                  |  1 +
 gcc/fortran/resolve.c                  |  9 +++-
 gcc/testsuite/gfortran.dg/pr102332.f90 | 69 ++++++++++++++++++++++++++
 5 files changed, 81 insertions(+), 4 deletions(-)
 create mode 100644 gcc/testsuite/gfortran.dg/pr102332.f90

diff --git a/gcc/fortran/expr.c b/gcc/fortran/expr.c
index b874607db1d..c1258e0eb06 100644
--- a/gcc/fortran/expr.c
+++ b/gcc/fortran/expr.c
@@ -5166,7 +5166,8 @@  gfc_get_variable_expr (gfc_symtree *var)

   if (var->n.sym->attr.flavor != FL_PROCEDURE
       && ((var->n.sym->as != NULL && var->n.sym->ts.type != BT_CLASS)
-	   || (var->n.sym->ts.type == BT_CLASS && CLASS_DATA (var->n.sym)
+	   || (var->n.sym->ts.type == BT_CLASS && var->n.sym->ts.u.derived
+	       && CLASS_DATA (var->n.sym)
 	       && CLASS_DATA (var->n.sym)->as)))
     {
       e->rank = var->n.sym->ts.type == BT_CLASS
diff --git a/gcc/fortran/match.c b/gcc/fortran/match.c
index 617fb35c9cd..41faa53b97a 100644
--- a/gcc/fortran/match.c
+++ b/gcc/fortran/match.c
@@ -6363,7 +6363,8 @@  select_type_set_tmp (gfc_typespec *ts)
       sym = tmp->n.sym;
       gfc_add_type (sym, ts, NULL);

-      if (selector->ts.type == BT_CLASS && selector->attr.class_ok)
+      if (selector->ts.type == BT_CLASS && selector->attr.class_ok
+	  && selector->ts.u.derived && CLASS_DATA (selector))
 	{
 	  sym->attr.pointer
 		= CLASS_DATA (selector)->attr.class_pointer;
diff --git a/gcc/fortran/primary.c b/gcc/fortran/primary.c
index d873264a08e..1f63028d179 100644
--- a/gcc/fortran/primary.c
+++ b/gcc/fortran/primary.c
@@ -2151,6 +2151,7 @@  gfc_match_varspec (gfc_expr *primary, int equiv_flag, bool sub_flag,
 	  && !(gfc_matching_procptr_assignment
 	       && sym->attr.flavor == FL_PROCEDURE))
       || (sym->ts.type == BT_CLASS && sym->attr.class_ok
+	  && sym->ts.u.derived && CLASS_DATA (sym)
 	  && (CLASS_DATA (sym)->attr.dimension
 	      || CLASS_DATA (sym)->attr.codimension)))
     {
diff --git a/gcc/fortran/resolve.c b/gcc/fortran/resolve.c
index bff1b35446f..591e8186007 100644
--- a/gcc/fortran/resolve.c
+++ b/gcc/fortran/resolve.c
@@ -5736,6 +5736,8 @@  resolve_variable (gfc_expr *e)
      can't be translated that way.  */
   if (sym->assoc && e->rank == 0 && e->ref && sym->ts.type == BT_CLASS
       && sym->assoc->target && sym->assoc->target->ts.type == BT_CLASS
+      && sym->assoc->target->ts.u.derived
+      && CLASS_DATA (sym->assoc->target)
       && CLASS_DATA (sym->assoc->target)->as)
     {
       gfc_ref *ref = e->ref;
@@ -5799,7 +5801,8 @@  resolve_variable (gfc_expr *e)
   /* Like above, but for class types, where the checking whether an array
      ref is present is more complicated.  Furthermore make sure not to add
      the full array ref to _vptr or _len refs.  */
-  if (sym->assoc && sym->ts.type == BT_CLASS
+  if (sym->assoc && sym->ts.type == BT_CLASS && sym->ts.u.derived
+      && CLASS_DATA (sym)
       && CLASS_DATA (sym)->attr.dimension
       && (e->ts.type != BT_DERIVED || !e->ts.u.derived->attr.vtype))
     {
@@ -9432,6 +9435,7 @@  resolve_select_type (gfc_code *code, gfc_namespace *old_ns)

       /* Check F03:C815.  */
       if ((c->ts.type == BT_DERIVED || c->ts.type == BT_CLASS)
+	  && selector_type
 	  && !selector_type->attr.unlimited_polymorphic
 	  && !gfc_type_is_extensible (c->ts.u.derived))
 	{
@@ -9442,7 +9446,8 @@  resolve_select_type (gfc_code *code, gfc_namespace *old_ns)
 	}

       /* Check F03:C816.  */
-      if (c->ts.type != BT_UNKNOWN && !selector_type->attr.unlimited_polymorphic
+      if (c->ts.type != BT_UNKNOWN
+	  && selector_type && !selector_type->attr.unlimited_polymorphic
 	  && ((c->ts.type != BT_DERIVED && c->ts.type != BT_CLASS)
 	      || !gfc_type_is_extension_of (selector_type, c->ts.u.derived)))
 	{
diff --git a/gcc/testsuite/gfortran.dg/pr102332.f90 b/gcc/testsuite/gfortran.dg/pr102332.f90
new file mode 100644
index 00000000000..f9557094083
--- /dev/null
+++ b/gcc/testsuite/gfortran.dg/pr102332.f90
@@ -0,0 +1,69 @@ 
+! { dg-do compile }
+! PR fortran/102332 - ICE in select_type_set_tmp
+! Contributed by G.Steinmetz
+
+program p
+  type t
+     real :: a, b
+  end type
+  class(t), allocatable :: x ! Valid
+  select type (y => x)
+  type is (t)
+     y%a = 0
+  end select
+end
+
+subroutine s0 (x)
+  type t
+     real :: a, b
+  end type
+  class(t) :: x ! Valid
+  select type (y => x)
+  type is (t)
+     y%a = 0
+  end select
+end
+
+subroutine s1
+  type t
+     real :: a, b
+  end type
+  class(t) :: x         ! { dg-error "must be dummy, allocatable or pointer" }
+  select type (y => x)
+  type is (t)
+     y%a = 0
+  end select
+end
+
+subroutine s3
+  type t
+     real :: a, b
+  end type
+  class(t) :: x         ! { dg-error "must be dummy, allocatable or pointer" }
+  select type (y => x)
+  class is (t)
+     y%a = 0
+  end select
+end
+
+subroutine s2
+  type t
+     real :: a, b
+  end type
+  class(t) :: x         ! { dg-error "must be dummy, allocatable or pointer" }
+  select type (y => x)
+  type default          ! { dg-error "Expected" }
+     y%a = 0
+  end select
+end
+
+subroutine s4
+  type t
+     real :: a, b
+  end type
+  class(t) :: x         ! { dg-error "must be dummy, allocatable or pointer" }
+  select type (y => x)
+  class default
+     y%a = 0
+  end select
+end
--
2.26.2