When update_address_taken rewrites a _Complex into SSA it changes
stores to real/imaginary parts to loads of the other component and
a COMPLEX_EXPR. That matches what gimplification does but it misses
suppression of diagnostics for the load of the other component.
The following patch adds that, syncing up gimplification and
update_address_taken behavior.
Bootstrapped and tested on x86_64-unknown-linux-gnu, pushed.
2022-03-31 Richard Biener <rguent...@suse.de>
PR tree-optimization/105109
* tree-ssa.cc (execute_update_addresses_taken): Suppress
diagnostics on the load of the other complex component.
* gcc.dg/uninit-pr105109.c: New testcase.
---
gcc/testsuite/gcc.dg/uninit-pr105109.c | 15 +++++++++++++++
gcc/tree-ssa.cc | 1 +
2 files changed, 16 insertions(+)
create mode 100644 gcc/testsuite/gcc.dg/uninit-pr105109.c
diff --git a/gcc/testsuite/gcc.dg/uninit-pr105109.c
b/gcc/testsuite/gcc.dg/uninit-pr105109.c
new file mode 100644
index 00000000000..001003ca261
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/uninit-pr105109.c
@@ -0,0 +1,15 @@
+/* { dg-do compile } */
+/* { dg-options "-O -Wuninitialized" } */
+
+static void foo(int dim,float _Complex f0[])
+{
+ int d;
+ f0[0] -= 3.14; /* { dg-bogus "uninitialized" } */
+ for (d = 0; d < dim; ++d) f0[0] += 3.14;
+}
+void bar(int dim, const float _Complex u_t[], float _Complex f0[])
+{
+ float _Complex exp[1] = {0.};
+ foo(dim, exp);
+ f0[0] = u_t[0] - exp[0];
+}
diff --git a/gcc/tree-ssa.cc b/gcc/tree-ssa.cc
index 6dcb3142869..a362a0a9ea6 100644
--- a/gcc/tree-ssa.cc
+++ b/gcc/tree-ssa.cc
@@ -1905,6 +1905,7 @@ execute_update_addresses_taken (void)
? REALPART_EXPR : IMAGPART_EXPR,
TREE_TYPE (other),
TREE_OPERAND (lhs, 0));
+ suppress_warning (lrhs);
gimple *load = gimple_build_assign (other, lrhs);
location_t loc = gimple_location (stmt);
gimple_set_location (load, loc);
--
2.34.1
