My sincere apologies to everyone, but especially Andrew Pinski
who warned me in advance that TRULY_NOOP_TRUNCATION results in
different code paths/optimizations on some targets. This restores
the build on nvptx-none (and presumably others) where mysteriously
(truncate:QI (reg:QI)) fails to be simplified to (reg:QI), which
is expected (everywhere) in my recently added self-tests.
[I'd even tested the functional change of the previous patch on
nvptx-none, but it's the new self-tests that I added for submission
that have been shown to be less portable across targets than hoped].
This patch has been committed as obvious, after testing that it
restores the build on nvptx-none and with a "make bootstrap" on
x86_64-pc-linux-gnu to guarantee no surprises [make -k check is
underway].
Sorry again.
2021-08-23 Roger Sayle <ro...@nextmovesoftware.com>
gcc/ChangeLog
* simplify-rtx.c (simplify_unary_operation_1): [TRUNCATE]:
Handle case where the operand is already the desired mode.
Roger
--
diff --git a/gcc/simplify-rtx.c b/gcc/simplify-rtx.c
index f3df614..8eea9fb 100644
--- a/gcc/simplify-rtx.c
+++ b/gcc/simplify-rtx.c
@@ -1268,6 +1268,9 @@ simplify_context::simplify_unary_operation_1 (rtx_code
code, machine_mode mode,
return temp;
}
+ /* Check for useless truncation. */
+ if (GET_MODE (op) == mode)
+ return op;
break;
case FLOAT_TRUNCATE:
