[llvm] r233693 - [AArch64] Add v8.1a "Rounding Double Multiply Add/Subtract" extension

[Tim Northover]

On 31 March 2015 at 09:14, Vladimir Sukharev <Vladimir.Sukharev at arm.com> wrote:
> Hi, that's very strange picture - http://lab.llvm.org:8080/green/job/clang-stage1-cmake-RA-incremental_check/6923/testReport/junit/LLVM/CodeGen_AArch64/arm64_neon_v8_1a_ll/
> This test is passed on my side, and it's weird I see apple syntax " sqrdmulh.4h v1, v1, v2", while "CHECK-V8a" denotes there was compilation without " -aarch64-neon-syntax=apple" flag.

I think the default syntax is inferred partially from the triple. The
easiest fix is to just add an explicit neon syntax (or probably a full
-mtriple).

Tim.