This is indeed a problem. Good catch; thanks!
-DeLesley
On Fri, Jul 1, 2011 at 11:50 AM, Martin Jambor <mjam...@suse.cz> wrote:
> Hi,
>
> On Fri, Jul 01, 2011 at 09:31:30AM -0700, Delesley Hutchins wrote:
>> > Just out of curiosity, I does your
>> > analysis crash also on the testcase below (add the Mutex field if it
>> > is necessary)?
>>
>> No, the static type of b (in b->test()) remains Bar* in gimple, so the
>> analysis works fine (i.e. the analyzer will not crash).
>>
>
> Sorry, I've had a look only just now and you do the analysis before
> early inlining and as long as you do that, the type is really Bar and
> you are fine as far as far as placement new is concerned.
>
> However, on the second thought, I still think you need to handle the
> case when BINFO_VIRTUALS are NULL in cp_get_virtual_function_decl
> (BTW, there is a non-langhook variant in gimple-fold.c called
> gimple_get_virt_method_for_binfo) because of the case below.
>
> Martin
>
>
> // { dg-do compile }
> // { dg-options "-Wthread-safety -O" }
> //#include "thread_annot_common.h"
>
> class Anc {
> public:
> int a;
> };
>
> class Foo : public Anc {
> public:
> // Mutex m;
>
> Foo();
> virtual ~Foo();
> virtual void doSomething() = 0;// EXCLUSIVE_LOCKS_REQUIRED(m) = 0;
> };
>
>
> class FooDerived : public Foo {
> public:
> FooDerived();
> virtual ~FooDerived();
> virtual void doSomething();
> };
>
> // reinterpret_cast
> void foo1(Anc* ptr)
> {
> reinterpret_cast<Foo*>(ptr)->doSomething();
> }
>
> // C-style cast
> void foo2(Anc* buf)
> {
> ((Foo*) buf)->doSomething();
> }
>
>
>
--
DeLesley Hutchins | Software Engineer | deles...@google.com | 505-206-0315
