shithub: libvpx

Info  •  Files  •  Log  •  Branches

Be explicit about overflow in vpx_variance16x16_sse2.

The product always fits in uint32_t, but the operands don't.

An optimizing compiler should generate the wraparound code.
(Verified with clang).

Change-Id: I25eb64df99152992bc898b8ccbb01d55c8d16e3c

--- a/vpx_dsp/x86/variance_sse2.c

+++ b/vpx_dsp/x86/variance_sse2.c

@@ -223,7 +223,7 @@

                                     unsigned int *sse) {

   int sum;

   vpx_get16x16var_sse2(src, src_stride, ref, ref_stride, sse, &sum);

-  return *sse - (((unsigned int)sum * sum) >> 8);

+  return *sse - (((uint32_t)((int64_t)sum * sum)) >> 8);

 }

 unsigned int vpx_variance32x32_sse2(const uint8_t *src, int src_stride,