The vp8_post_proc_down_and_across_mb_row_sse2() needs space for an
even number of macroblocks, as they are read two at a time for the
chroma planes. Round up the width during the allocation of
pp_limits_buffer to support this.
Change-Id: Ibfc10c7be290d961ab23ac3dde12a7bb96c12af0
vpx_memset(oci->post_proc_buffer.buffer_alloc, 128,
oci->post_proc_buffer.frame_size);
- /* Allocate buffer to store post-processing filter coefficients. */
- oci->pp_limits_buffer = vpx_memalign(16, 24 * oci->mb_cols);
+ /* Allocate buffer to store post-processing filter coefficients.
+ *
+ * Note: Round up mb_cols to support SIMD reads
+ */
+ oci->pp_limits_buffer = vpx_memalign(16, 24 * ((oci->mb_cols + 1) & ~1));
if (!oci->pp_limits_buffer)
goto allocation_fail;
#endif