Remove checking for a case that (most probably) never happens as the
only expected instruction is OP_PHISOURCE.
For the sake of assurance, assert added with a comment easing the fix if
the assumption of OP_PHISOURCE is wrong (better then masking a problem).
Also, use `def' as existing shortcut for `phi->def'.
Tested the same way as the previous patch (together).
Signed-off-by: Jan Pokorny <pokorny_jan@xxxxxxxxx>
---
liveness.c | 8 +++-----
1 files changed, 3 insertions(+), 5 deletions(-)
diff --git a/liveness.c b/liveness.c
index eeff0f7..02ceed9 100644
--- a/liveness.c
+++ b/liveness.c
@@ -23,11 +23,9 @@ static void phi_defines(struct instruction * phi_node, pseudo_t target,
def = phi->def;
if (!def || !def->bb)
continue;
- if (def->opcode == OP_PHI) {
- phi_defines(def, target, defines);
- continue;
- }
- defines(def->bb, phi->def, target);
+ /* if this ever fails (very unlikely), it's a sign of a regression */
+ assert(def->opcode == OP_PHISOURCE);
+ defines(def->bb, def, target);
} END_FOR_EACH_PTR(phi);
}
--
1.7.1
--
To unsubscribe from this list: send the line "unsubscribe linux-sparse" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
