[llvm] [NVPTX] Fix 64 bits rotations with large shift values (PR #89399)

[Nicolas Miller via llvm-commits]

================
@@ -58,3 +58,46 @@ define i32 @rotl0(i32 %x) {
   %t2 = or i32 %t0, %t1
   ret i32 %t2
 }
+
+declare i64 @llvm.fshl.i64(i64, i64, i64)
+declare i64 @llvm.fshr.i64(i64, i64, i64)
+
+; SM35: rotl64
+define i64 @rotl64(i64 %a, i64 %n) {
----------------
npmiller wrote:

Oh right! Updated with the script now, it was confusing for a little bit because it doesn't understand `-march` so I had to swap to `--mtriple`.

https://github.com/llvm/llvm-project/pull/89399