diff options
Diffstat (limited to 'src/dsp/arm/inverse_transform_neon.cc')
| -rw-r--r-- | src/dsp/arm/inverse_transform_neon.cc | 9 |
1 files changed, 5 insertions, 4 deletions
diff --git a/src/dsp/arm/inverse_transform_neon.cc b/src/dsp/arm/inverse_transform_neon.cc index 452f14a..cc4e4a4 100644 --- a/src/dsp/arm/inverse_transform_neon.cc +++ b/src/dsp/arm/inverse_transform_neon.cc @@ -345,11 +345,12 @@ LIBGAV1_ALWAYS_INLINE void ButterflyRotation_FirstIsZero(int16x8_t* a, int16x8_t* b, const int angle, const bool flip) { + // Clang < 14 targeting armv8.1-a+ optimizes vqrdmulhq_n_s16 and vqsubq_s16 + // (in HadamardRotation) into vqrdmlshq_s16 resulting in an "off by one" + // error. This behavior was fixed in 14.0.0: + // https://github.com/llvm/llvm-project/commit/82973edfb72a95b442fa6d2bb404e15a4031855e #if defined(__ARM_FEATURE_QRDMX) && defined(__aarch64__) && \ - defined(__clang__) // ARM v8.1-A - // Clang optimizes vqrdmulhq_n_s16 and vqsubq_s16 (in HadamardRotation) into - // vqrdmlshq_s16 resulting in an "off by one" error. For now, do not use - // vqrdmulhq_n_s16(). + defined(__clang__) && __clang_major__ < 14 const int16_t cos128 = Cos128(angle); const int16_t sin128 = Sin128(angle); const int32x4_t x0 = vmull_n_s16(vget_low_s16(*b), -sin128); |