[Bug middle-end/93764] [10 Regression] lzo 2.10 test suite fails with -O2 and -O3

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93764

--- Comment #5 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
The first difference is during DSE1:
--- lzo_init2.i.036t.dse1_      2020-02-17 07:36:46.000000000 -0500
+++ lzo_init2.i.036t.dse1       2020-02-17 07:37:08.000000000 -0500
@@ -109,7 +109,6 @@ lzo_config_check ()
   _3 = _2 == 0;
   # RANGE [0, 1] NONZERO 1
   _4 = (unsigned int) _3;
-  u.b[0] = 128;
   _6 = MEM[(long unsigned int *)p_58];
   _7 = _6 == 128;
   # RANGE [0, 1] NONZERO 1