c: Fix up -Wunused-but-set-* warnings for _Atomics [PR99588]
authorJakub Jelinek <jakub@redhat.com>
Fri, 19 Mar 2021 21:54:31 +0000 (22:54 +0100)
committerJakub Jelinek <jakub@redhat.com>
Fri, 19 Mar 2021 21:54:31 +0000 (22:54 +0100)
commit3279a9a5a9a2e4e17175678cb7b15613495e306e
tree1874979bd0e80e0189c05bfa79b2fe94625ce9b6
parente5d74554b5759637439c7cab00903f864653af5f
c: Fix up -Wunused-but-set-* warnings for _Atomics [PR99588]

As the following testcases show, compared to -D_Atomic= case we have many
-Wunused-but-set-* warning false positives.
When an _Atomic variable/parameter is read, we call mark_exp_read on it in
convert_lvalue_to_rvalue, but build_atomic_assign does not.
For consistency with the non-_Atomic case where we mark_exp_read the lhs
for lhs op= ... but not for lhs = ..., this patch does that too.
But furthermore we need to pattern match the trees emitted by _Atomic store,
so that _Atomic store itself is not marked as being a variable read, but
when the result of the store is used, we mark it.

2021-03-19  Jakub Jelinek  <jakub@redhat.com>

PR c/99588
* c-typeck.c (mark_exp_read): Recognize what build_atomic_assign
with modifycode NOP_EXPR produces and mark the _Atomic var as read
if found.
(build_atomic_assign): For modifycode of NOP_EXPR, use COMPOUND_EXPRs
rather than STATEMENT_LIST.  Otherwise call mark_exp_read on lhs.
Set TREE_SIDE_EFFECTS on the TARGET_EXPR.

* gcc.dg/Wunused-var-5.c: New test.
* gcc.dg/Wunused-var-6.c: New test.
gcc/c/c-typeck.c
gcc/testsuite/gcc.dg/Wunused-var-5.c [new file with mode: 0644]
gcc/testsuite/gcc.dg/Wunused-var-6.c [new file with mode: 0644]