================
@@ -582,6 +582,11 @@ TARGET_BUILTIN(__builtin_ppc_bcdsub_p, "iiV16UcV16Uc", "",
// P9 Binary-coded decimal (BCD) builtins.
TARGET_BUILTIN(__builtin_ppc_bcdcopysign, "V16UcV16UcV16Uc", "",
"power9-vector")
TARGET_BUILTIN(__builtin_ppc_bcdsetsign, "V16UcV16UcUc", "t", "power9-vector")
+TARGET_BUILTIN(__builtin_ppc_bcdshift, "V16UcV16UciUc", "t", "power9-vector")
+TARGET_BUILTIN(__builtin_ppc_bcdshiftround, "V16UcV16UciUc", "t",
"power9-vector")
+TARGET_BUILTIN(__builtin_ppc_bcdtruncate, "V16UcV16UciUc", "t",
"power9-vector")
----------------
RolandF77 wrote:
Tried an example with an invalid first parameter type to __builtin_ppc_bcdshift
and got the error message:
bcd.c:35:16: error: expected a 'float', 'double' or '__float128' for the first
argument
The first parameter is a supposed to be a vector. You are using the diagnostic
from the test data class builtin and that is inaccurate. I suspect you need to
add a diagnostic.
https://github.com/llvm/llvm-project/pull/154715
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
