This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [tuples][patch] Convert pass_object_sizes to tuples


On Tue, Mar 18, 2008 at 14:04, Bill Maddox <maddox@google.com> wrote:
>
> On Tue, Mar 18, 2008 at 10:57 AM, Diego Novillo <dnovillo@google.com> wrote:
>   > On Tue, Mar 18, 2008 at 13:53, Bill Maddox <maddox@google.com> wrote:
>   >  > On Tue, Mar 18, 2008 at 6:20 AM, Jakub Jelinek <jakub@redhat.com> wrote:
>   >  >
>   >  >  >  Both of these __bos calls should be folded to 64.
>   >  >  >  compute_builtin_object_size is called by builtins.c folding, so the
>   >  >  >  arguments aren't necessarily GIMPLE, can be GENERIC.
>   >  >
>   >  >  Hmm.  This could be a problem.  The patch handles calls only when
>   >  >  presented as GIMPLE_CALL statements.
>   >
>   >  And in fact we should fail much earlier because the gimple builders
>   >  will not allow a CALL_EXPR in the argument list for a GIMPLE_CALL.
>
>
>
>  If we can call compute_builtin_object size on GENERIC, however, it looks
>   like I'll have to put the handler for CALL_EXPR back in.  Likewise for
>   POINTER_PLUS_EXPR.

Hmm, no.  I don't think this is possible now.  We call
compute_builtin_object_size mostly from fold_builtin_object_size,
which is only invoked during GIMPLE optimization.

If it is possible to call this code path from GENERIC, we should have
a Good Reason.  If not, I would like to block this possibility.


Diego.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]