[PATCH] D26092: Implement vector shift builtins - clang portion.

[Kit Barton via llvm-commits]

kbarton accepted this revision.
kbarton added a comment.
This revision is now accepted and ready to land.

Aside from some formatting issues, this LGTM.



================
Comment at: lib/Headers/altivec.h:7591
+                                                             vector bool long long  __b,
+                                                             unsigned const int __c) {
+  unsigned char __d = __c & 0x0F;
----------------
Lines too long. Please run clang-format on these sections you've added to adjust the formatting. 


================
Comment at: lib/Headers/altivec.h:7702
+
+static __inline__ vector signed short __ATTRS_o_ai vec_sldw(
+    vector signed short __a, vector signed short __b, unsigned const int __c) {
----------------
This is a minor nit, but I seem to remember definitions in this file being orderd by size: char, short, int, long, long long.
Can you please move these definitions up, to between the char and int, for consistency?


================
Comment at: lib/Headers/altivec.h:7713
+
+#ifdef __POWER9_VECTOR__
+static __inline__ vector unsigned char __ATTRS_o_ai
----------------
Add /* vec_slv */ for consistency


https://reviews.llvm.org/D26092