arm64: clang assembler compatability

dl/sp/api/armSP.h

 /*
  *  Copyright (c) 2013 The WebRTC project authors. All Rights Reserved.
  *
  *  Use of this source code is governed by a BSD-style license
  *  that can be found in the LICENSE file in the root of the source
  *  tree. An additional intellectual property rights grant can be found
  *  in the file PATENTS.  All contributing project authors may
  *  be found in the AUTHORS file in the root of the source tree.
  *
  *  This file was originally licensed as follows. It has been
@@ skipping 75 matching lines @@
 {
     OMX_U32 N;
     OMX_U16* pBitRev;
     OMX_FC32* pTwiddle;
     OMX_FC32* pBuf;
 } ARMsFFTSpec_FC32;
 
 /*
  * Compute log2(x), where x must be a power of 2.
  */
-static inline OMX_U32 fastlog2(long x) {
+static inline OMX_U32 fastlog2(OMX_U32 x) {

[Raymond Toy (Google), 2015/10/30 22:38:46]

Why the change from long to OMX_U32?

[Riku Voipio, 2015/11/02 12:24:12]

clang throws a warning:
../../third_party/openmax_dl/dl/sp/api/armSP.h:101:14: warning: value size does
not match register size specified by
 the constraint and modifier [-Wasm-operand-widths]
      : "=r"(out)

The only users of this fastlog2 call it with OMX_U32 value (spec->N). So instead
of changing the just the out value to long, I changed the whole function to
32bit. 

I'll update the patch to only change the return type.

   OMX_U32 out;
-  asm("clz %0,%1\n\t"
-      "sub %0, %0, #63\n\t"
-      "neg %0, %0\n\t"
+  asm("clz %w0,%w1\n\t"
+      "sub %w0, %w0, #63\n\t"

[Raymond Toy (Google), 2015/10/30 22:38:46]

If %w0 means a 32-bit register than clz is probably only returning a max value
of 32, so subtracting 63 is probably wrong.

[Riku Voipio, 2015/11/02 12:24:12]

Will be fixed in the next iteration by keeping the function 64bit.

+      "neg %w0, %w0\n\t"
       : "=r"(out)
       : "r"(x)
       :);
   return out;
 }
 
 /*
  *  Validate args. All pointers must be non-NULL; the source and
  *  destination pointers must be aligned on a 32-byte boundary; the
  *  FFT spec must have non-NULL pointers; and the FFT size must be
@@ skipping 18 matching lines @@
 #ifdef __cplusplus
 }
 #endif
 
 #endif
 
 /*End of File*/