[Bug tree-optimization/115652] [15 Regression] GCN: FAIL: gcc.dg/vect/pr70138-{1,2}.c (internal compiler error: verify_ssa failed)

["rguenth at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115652

--- Comment #5 from Richard Biener <rguenth at gcc dot gnu.org> ---
(In reply to Jeffrey A. Law from comment #4)
> That patch seems to be causing ssa verification failures (should be visible
> with riscv64-elf cross):
> 
> make check-gcc RUNTESTFLAGS=rvv.exp=vsetvl_bug-2.c
> 
> Running
> /home/jlaw/jenkins/workspace/riscv64-elf/gcc/gcc/testsuite/gcc.target/riscv/
> rvv/rvv.exp ...
> FAIL: gcc.target/riscv/rvv/vsetvl/vsetvl_bug-2.c   -O0  (internal compiler
> error: verify_ssa failed)
> FAIL: gcc.target/riscv/rvv/vsetvl/vsetvl_bug-2.c   -O0  (test for excess
> errors)
> FAIL: gcc.target/riscv/rvv/vsetvl/vsetvl_bug-2.c   -O1  (internal compiler
> error: verify_ssa failed)
> FAIL: gcc.target/riscv/rvv/vsetvl/vsetvl_bug-2.c   -O1  (test for excess
> errors)
> FAIL: gcc.target/riscv/rvv/vsetvl/vsetvl_bug-2.c   -O2  (internal compiler
> error: verify_ssa failed)
> FAIL: gcc.target/riscv/rvv/vsetvl/vsetvl_bug-2.c   -O2  (test for excess
> errors)
> FAIL: gcc.target/riscv/rvv/vsetvl/vsetvl_bug-2.c   -O3 -fomit-frame-pointer
> -funroll-loops -fpeel-loops -ftracer -finline-functions  (internal compiler
> error: verify_ssa failed)
> FAIL: gcc.target/riscv/rvv/vsetvl/vsetvl_bug-2.c   -O3 -fomit-frame-pointer
> -funroll-loops -fpeel-loops -ftracer -finline-functions  (test for excess
> errors)
> FAIL: gcc.target/riscv/rvv/vsetvl/vsetvl_bug-2.c   -O3 -g  (internal
> compiler error: verify_ssa failed)
> FAIL: gcc.target/riscv/rvv/vsetvl/vsetvl_bug-2.c   -O3 -g  (test for excess
> errors)
> FAIL: gcc.target/riscv/rvv/vsetvl/vsetvl_bug-2.c   -Os  (internal compiler
> error: verify_ssa failed)
> FAIL: gcc.target/riscv/rvv/vsetvl/vsetvl_bug-2.c   -Os  (test for excess
> errors)
> 
> There's a handful of other failures since yesterday, but I suspect they're
> all the same root cause.

I can reproduce the above, we get

  vect__85.21_1 = .COND_LEN_XOR ({ -1, ... }, vect_l_437__lsm.13_110.20_2,
vect__87.19_22, vect_l_437__lsm.13_110.20_2, loop_len_54, 0);
  vect__87.19_22 = (vector([32,32]) int) vect__90.18_9;

I'm testing a fix:

diff --git a/gcc/tree-vect-slp.cc b/gcc/tree-vect-slp.cc
index 896ea6dc14c..ce214dd652e 100644
--- a/gcc/tree-vect-slp.cc
+++ b/gcc/tree-vect-slp.cc
@@ -9752,7 +9752,8 @@ vect_schedule_slp_node (vec_info *vinfo,
              {
                gimple_stmt_iterator si2
                  = gsi_after_labels (LOOP_VINFO_LOOP (loop_vinfo)->header);
-               if (vect_stmt_dominates_stmt_p (last_stmt, *si2))
+               if (last_stmt != *si2
+                   && vect_stmt_dominates_stmt_p (last_stmt, *si2))
                  si = si2;
              }
        }