On 9/23/07, Manuel López-Ibáñez <[EMAIL PROTECTED]> wrote:
> On 23/09/2007, Richard Guenther <[EMAIL PROTECTED]> wrote:
> > On 9/23/07, Manuel López-Ibáñez <[EMAIL PROTECTED]> wrote:
> > > On 23/09/2007, Richard Guenther <[EMAIL PROTECTED]> wrote:
> > > > On 9/23/07, Manuel López-Ibáñez <[EMAIL PROTECTED]> wrote:
> > > > >
> > > > > Am I wrong? I have the same problem when I try to enable SSA at O0 and
> > > > > my current solution is the dummy pass.
> > > >
> > > > Right, a dummy pass after salias works.
> > > >
> > >
> > > And *before* salias? Does it make a difference? It suits me better for
> > > my purposes.
> >
> > No, after salias you need to run rebuild_alias. And for efficiency you'd
> > of course remove the TODO from salias itself.
> >
>
> I am not so concerned about efficiency ATM, I am trying to build SSA at O0.
>
> Anyway, it doesn't work, verify_ssa fails with
>
> /home/manuel/src/trunk/libgcc/../gcc/libgcc2.c: In function '__negti2':
> /home/manuel/src/trunk/libgcc/../gcc/libgcc2.c:74: error: found real
> variable when subvariables should have appeared
> while verifying SSA_NAME uu_12 in statement
> # SFT.21 = VDEF <SFT.21>
> # SFT.22 = VDEF <SFT.22>
> # SFT.23 = VDEF <SFT.23>
> uu = {};
>
> and
>
> /home/manuel/src/trunk/libgcc/../gcc/libgcc2.c:74: error: expected an
> SSA_NAME object
> /home/manuel/src/trunk/libgcc/../gcc/libgcc2.c:74: error: in statement
> # SFT.21 = VDEF <SFT.21>
> # SFT.22 = VDEF <SFT.22>
> # SFT.23 = VDEF <SFT.23> { SFT.21 SFT.22 SFT.23 }
> uu = {};
>
> :-(
salias never updated ssa, it expected rebuild_alias to do it.
There is no point in updating ssa after salias but before building
alias info, and building alias info doesn't care about vuse/vdef's
being in ssa form.
So just remove the verify ssa.
