From: James Zern <jzern@google.com>
Date: Sat, 6 Nov 2021 23:43:11 +0000 (-0700)
Subject: mem_sse2.h: storeu_uint32 -> storeu_int32
X-Git-Tag: v1.12.0-rc1~124^2~1
X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=2e73da326a26ddc367317e860e24e274947c26d8;p=libvpx

mem_sse2.h: storeu_uint32 -> storeu_int32

this changes the parameter to int32_t which matches the type with usage
of this call using _mm_cvtsi128_si32() as a parameter. quiets an
implicit conversion warning with clang-11 -fsanitize=undefined

Change-Id: I1e9e9ffac5d2996962d29611458311221eca8ea0
---

diff --git a/vp8/common/x86/bilinear_filter_sse2.c b/vp8/common/x86/bilinear_filter_sse2.c
index 9bf65d804..ff6cbbd68 100644
--- a/vp8/common/x86/bilinear_filter_sse2.c
+++ b/vp8/common/x86/bilinear_filter_sse2.c
@@ -313,10 +313,10 @@ static INLINE void vertical_4x4(uint16_t *src, uint8_t *dst, const int stride,
       const __m128i compensated = _mm_add_epi16(sum, round_factor);
       const __m128i shifted = _mm_srai_epi16(compensated, VP8_FILTER_SHIFT);
       __m128i packed = _mm_packus_epi16(shifted, shifted);
-      storeu_uint32(dst, _mm_cvtsi128_si32(packed));
+      storeu_int32(dst, _mm_cvtsi128_si32(packed));
       packed = _mm_srli_si128(packed, 4);
       dst += stride;
-      storeu_uint32(dst, _mm_cvtsi128_si32(packed));
+      storeu_int32(dst, _mm_cvtsi128_si32(packed));
       dst += stride;
       src += 8;
     }
diff --git a/vpx_dsp/x86/loopfilter_sse2.c b/vpx_dsp/x86/loopfilter_sse2.c
index b6ff24834..347c9fdbe 100644
--- a/vpx_dsp/x86/loopfilter_sse2.c
+++ b/vpx_dsp/x86/loopfilter_sse2.c
@@ -211,21 +211,21 @@ void vpx_lpf_vertical_4_sse2(uint8_t *s, int pitch, const uint8_t *blimit,
   // 00 10 20 30 01 11 21 31  02 12 22 32 03 13 23 33
   ps1ps0 = _mm_unpacklo_epi8(ps1ps0, x0);
 
-  storeu_uint32(s + 0 * pitch - 2, _mm_cvtsi128_si32(ps1ps0));
+  storeu_int32(s + 0 * pitch - 2, _mm_cvtsi128_si32(ps1ps0));
   ps1ps0 = _mm_srli_si128(ps1ps0, 4);
-  storeu_uint32(s + 1 * pitch - 2, _mm_cvtsi128_si32(ps1ps0));
+  storeu_int32(s + 1 * pitch - 2, _mm_cvtsi128_si32(ps1ps0));
   ps1ps0 = _mm_srli_si128(ps1ps0, 4);
-  storeu_uint32(s + 2 * pitch - 2, _mm_cvtsi128_si32(ps1ps0));
+  storeu_int32(s + 2 * pitch - 2, _mm_cvtsi128_si32(ps1ps0));
   ps1ps0 = _mm_srli_si128(ps1ps0, 4);
-  storeu_uint32(s + 3 * pitch - 2, _mm_cvtsi128_si32(ps1ps0));
+  storeu_int32(s + 3 * pitch - 2, _mm_cvtsi128_si32(ps1ps0));
 
-  storeu_uint32(s + 4 * pitch - 2, _mm_cvtsi128_si32(qs1qs0));
+  storeu_int32(s + 4 * pitch - 2, _mm_cvtsi128_si32(qs1qs0));
   qs1qs0 = _mm_srli_si128(qs1qs0, 4);
-  storeu_uint32(s + 5 * pitch - 2, _mm_cvtsi128_si32(qs1qs0));
+  storeu_int32(s + 5 * pitch - 2, _mm_cvtsi128_si32(qs1qs0));
   qs1qs0 = _mm_srli_si128(qs1qs0, 4);
-  storeu_uint32(s + 6 * pitch - 2, _mm_cvtsi128_si32(qs1qs0));
+  storeu_int32(s + 6 * pitch - 2, _mm_cvtsi128_si32(qs1qs0));
   qs1qs0 = _mm_srli_si128(qs1qs0, 4);
-  storeu_uint32(s + 7 * pitch - 2, _mm_cvtsi128_si32(qs1qs0));
+  storeu_int32(s + 7 * pitch - 2, _mm_cvtsi128_si32(qs1qs0));
 }
 
 void vpx_lpf_horizontal_16_sse2(unsigned char *s, int pitch,
diff --git a/vpx_dsp/x86/mem_sse2.h b/vpx_dsp/x86/mem_sse2.h
index 258ab38e6..75fa2b0b7 100644
--- a/vpx_dsp/x86/mem_sse2.h
+++ b/vpx_dsp/x86/mem_sse2.h
@@ -16,7 +16,7 @@
 
 #include "./vpx_config.h"
 
-static INLINE void storeu_uint32(void *dst, uint32_t v) {
+static INLINE void storeu_int32(void *dst, int32_t v) {
   memcpy(dst, &v, sizeof(v));
 }