On Thu, Nov 7, 2024 at 12:50 AM Richard Biener
<richard.guent...@gmail.com> wrote:
>
> On Thu, Nov 7, 2024 at 12:43 AM Andrew Pinski <quic_apin...@quicinc.com>
> wrote:
> >
> > After the last patch, we also want to record `(A CMP B) != 0`
> > as `(A CMP B)` and `(A CMP B) == 0` as `(A CMP B)` with the
> > true/false edges swapped.
> >
> > This shows up more due to the new handling of
> > `(A | B) ==/!= 0` in insert_predicates_for_cond
> > as now we can notice these comparisons which were not seen before.
> >
> > This is enough to fix the original issue in `gcc.dg/tree-ssa/pr111456-1.c`
> > and make sure we don't regress it when enhancing ifcombine.
> >
> > This adds that predicate and allows us to optimize f
> > in fre-predicated-3.c.
> >
> > Changes since v1:
> > * v2: Use vn_valueize.
> >
> > Bootstrapped and tested on x86_64-linux-gnu.
> >
> > PR tree-optimization/117414
> >
> > gcc/ChangeLog:
> >
> > * tree-ssa-sccvn.cc (insert_predicates_for_cond): Handle `(A CMP B)
> > !=/== 0`.
> >
> > gcc/testsuite/ChangeLog:
> >
> > * gcc.dg/tree-ssa/fre-predicated-3.c: New test.
> >
> > Signed-off-by: Andrew Pinski <quic_apin...@quicinc.com>
> > ---
> > .../gcc.dg/tree-ssa/fre-predicated-3.c | 46 +++++++++++++++++++
> > gcc/tree-ssa-sccvn.cc | 14 ++++++
> > 2 files changed, 60 insertions(+)
> > create mode 100644 gcc/testsuite/gcc.dg/tree-ssa/fre-predicated-3.c
> >
> > diff --git a/gcc/testsuite/gcc.dg/tree-ssa/fre-predicated-3.c
> > b/gcc/testsuite/gcc.dg/tree-ssa/fre-predicated-3.c
> > new file mode 100644
> > index 00000000000..4a89372fd70
> > --- /dev/null
> > +++ b/gcc/testsuite/gcc.dg/tree-ssa/fre-predicated-3.c
> > @@ -0,0 +1,46 @@
> > +/* { dg-do compile } */
> > +/* { dg-options "-O2 -fdump-tree-optimized" } */
> > +
> > +/* PR tree-optimization/117414 */
> > +
> > +/* Fre1 should figure out that `*aaa != 0`
> > + For f0, f1, and f2. */
> > +
> > +void foo();
> > +int f(int *aaa, int j, int t)
> > +{
> > + int b = *aaa;
> > + int c = b == 0;
> > + int d = t != 1;
> > + if (c | d)
> > + return 0;
> > +
> > + for(int i = 0; i < j; i++)
> > + {
> > + if (*aaa)
> > + ;
> > + else
> > + foo();
> > + }
> > + return 0;
> > +}
> > +
> > +int f1(int *aaa, int j, int t)
> > +{
> > + int b = *aaa;
> > + if (b == 0)
> > + return 0;
> > + if (t != 1)
> > + return 0;
> > + for(int i = 0; i < j; i++)
> > + {
> > + if (*aaa)
> > + ;
> > + else
> > + foo();
> > + }
> > + return 0;
> > +}
> > +
> > +/* { dg-final { scan-tree-dump-not "foo " "optimized" } } */
> > +/* { dg-final { scan-tree-dump "return 0;" "optimized" } } */
> > diff --git a/gcc/tree-ssa-sccvn.cc b/gcc/tree-ssa-sccvn.cc
> > index c6dddd0ba6d..67ed2cd8ffe 100644
> > --- a/gcc/tree-ssa-sccvn.cc
> > +++ b/gcc/tree-ssa-sccvn.cc
> > @@ -7948,6 +7948,20 @@ insert_predicates_for_cond (tree_code code, tree
> > lhs, tree rhs,
> > && (code == NE_EXPR || code == EQ_EXPR))
> > {
> > gimple *def_stmt = SSA_NAME_DEF_STMT (lhs);
> > + /* (A CMP B) != 0 is the same as (A CMP B).
> > + (A CMP B) == 0 is just (A CMP B) with the edges swapped. */
> > + if (is_gimple_assign (def_stmt)
> > + && TREE_CODE_CLASS (gimple_assign_rhs_code (def_stmt)) ==
> > tcc_comparison)
> > + {
> > + tree_code nc = gimple_assign_rhs_code (def_stmt);
> > + tree nlhs = vn_valueize (gimple_assign_rhs1 (def_stmt));
> > + tree nrhs = vn_valueize (gimple_assign_rhs2 (def_stmt));
> > + edge nt = true_e;
> > + edge nf = false_e;
> > + if (code == EQ_EXPR)
> > + std::swap (nt, nf);
>
> I'll note there's canonicalization for tree_swap_operands as well,
> _1 < _2 vs. _2 > _1 - but I think this can be fixed as followup?
Yes I will handle that later today in a followup patch because it will
touch other areas of sccvn including places which does the lookup.
Thanks,
Andrew
>
> OK.
>
> Thanks,
> Richard.
>
> > + insert_predicates_for_cond (nc, nlhs, nrhs, nt, nf);
> > + }
> > /* (a | b) == 0 ->
> > on true edge assert: a == 0 & b == 0. */
> > /* (a | b) != 0 ->
> > --
> > 2.43.0
> >
