Hi Will,
On Tue, Sep 25, 2018 at 10:47:25AM -0500, Will Schmidt wrote:
> [v2 updates]
> Added the -longlong testcase changes.
> Add 'lxv' to expected instruction list (for power9 targets).
> Update and simplify some commentary sprinkled through the tests.
>
> OK for trunk?
It is still okay afaic
Hi,
This is an updated version of the previously posted patch.
This updates those powerpc testsuite tests that are using the
vec_splat() builtin with an invalid arg1. Per discussions during the
review of gimple-folding for vec_splat(), it was clarified
that arg1 for a vec_splat() call should
Hi Will,
On Thu, Sep 20, 2018 at 01:52:55PM -0500, Will Schmidt wrote:
> This updates those powerpc testsuite tests that are using the
> vec_splat() builtin with an invalid arg1. Per discussions during the
> review of gimple-folding for vec_splat(), it was clarified
> that arg1 for vec_splat()
Hi,
This updates those powerpc testsuite tests that are using the
vec_splat() builtin with an invalid arg1. Per discussions during the
review of gimple-folding for vec_splat(), it was clarified
that arg1 for vec_splat() should be a valid index into the
referenced vector (no modulo).
OK fo
