--- src/opts/SkRasterPipeline_opts.h.orig
+++ src/opts/SkRasterPipeline_opts.h
@@ -1793,10 +1793,17 @@
     *a = cast(wide & (15<< 0)) * (1.0f / (15<< 0));
 }
 SI void from_8888(U32 _8888, F* r, F* g, F* b, F* a) {
+#ifdef SK_CPU_BENDIAN
+    *r = cast((_8888 >> 24)       ) * (1/255.0f);
+    *g = cast((_8888 >> 16) & 0xff) * (1/255.0f);
+    *b = cast((_8888 >>  8) & 0xff) * (1/255.0f);
+    *a = cast((_8888      ) & 0xff) * (1/255.0f);
+#else
     *r = cast((_8888      ) & 0xff) * (1/255.0f);
     *g = cast((_8888 >>  8) & 0xff) * (1/255.0f);
     *b = cast((_8888 >> 16) & 0xff) * (1/255.0f);
     *a = cast((_8888 >> 24)       ) * (1/255.0f);
+#endif
 }
 SI void from_88(U16 _88, F* r, F* g) {
     U32 wide = expand(_88);
@@ -2383,10 +2390,17 @@
     auto ptr = ptr_at_xy<uint32_t>(ctx, dx,dy);
 
     U32 dst = load<U32>(ptr);
+#ifdef SK_CPU_BENDIAN
+    dr = cast((dst >> 24)       );
+    dg = cast((dst >> 16) & 0xff);
+    db = cast((dst >>  8) & 0xff);
+    da = cast((dst      ) & 0xff);
+#else
     dr = cast((dst      ) & 0xff);
     dg = cast((dst >>  8) & 0xff);
     db = cast((dst >> 16) & 0xff);
     da = cast((dst >> 24)       );
+#endif
     // {dr,dg,db,da} are in [0,255]
     // { r, g, b, a} are in [0,  1] (but may be out of gamut)
 
@@ -2397,10 +2411,17 @@
     // { r, g, b, a} are now in [0,255]  (but may be out of gamut)
 
     // to_unorm() clamps back to gamut.  Scaling by 1 since we're already 255-based.
+#ifdef SK_CPU_BENDIAN
+    dst = to_unorm(r, /*scale=*/1, /*bias=*/0.f, /*maxI=*/255) << 24
+        | to_unorm(g, /*scale=*/1, /*bias=*/0.f, /*maxI=*/255) << 16
+        | to_unorm(b, /*scale=*/1, /*bias=*/0.f, /*maxI=*/255) <<  8
+        | to_unorm(a, /*scale=*/1, /*bias=*/0.f, /*maxI=*/255);
+#else
     dst = to_unorm(r, /*scale=*/1, /*bias=*/0.f, /*maxI=*/255)
         | to_unorm(g, /*scale=*/1, /*bias=*/0.f, /*maxI=*/255) <<  8
         | to_unorm(b, /*scale=*/1, /*bias=*/0.f, /*maxI=*/255) << 16
         | to_unorm(a, /*scale=*/1, /*bias=*/0.f, /*maxI=*/255) << 24;
+#endif
     store(ptr, dst);
 }
 
@@ -2953,10 +2974,17 @@
 HIGHP_STAGE(store_8888, const SkRasterPipelineContexts::MemoryCtx* ctx) {
     auto ptr = ptr_at_xy<uint32_t>(ctx, dx,dy);
 
+#ifdef SK_CPU_BENDIAN
+    U32 px = to_unorm(r, 255) << 24
+           | to_unorm(g, 255) << 16
+           | to_unorm(b, 255) <<  8
+           | to_unorm(a, 255);
+#else
     U32 px = to_unorm(r, 255)
            | to_unorm(g, 255) <<  8
            | to_unorm(b, 255) << 16
            | to_unorm(a, 255) << 24;
+#endif
     store(ptr, px);
 }
 
@@ -6140,11 +6168,18 @@
     };
 #endif
 #if !defined(SKRP_CPU_LSX)
+#ifdef SK_CPU_BENDIAN
+    *r = cast_U16(rgba >>   16) >>  8;
+    *g = cast_U16(rgba >>   16) & 255;
+    *b = cast_U16(rgba & 65535) >>  8;
+    *a = cast_U16(rgba & 65535) & 255;
+#else
     *r = cast_U16(rgba & 65535) & 255;
     *g = cast_U16(rgba & 65535) >>  8;
     *b = cast_U16(rgba >>   16) & 255;
     *a = cast_U16(rgba >>   16) >>  8;
 #endif
+#endif
 }
 
 SI void load_8888_(const uint32_t* ptr, U16* r, U16* g, U16* b, U16* a) {
@@ -6197,10 +6232,15 @@
     }};
     vst4_u8((uint8_t*)(ptr), rgba);
 #else
+#ifdef SK_CPU_BENDIAN
+    store(ptr, cast<U32>((r<<8) | g) << 16
+             | cast<U32>((b<<8) | a) << 0);
+#else
     store(ptr, cast<U32>(r | (g<<8)) <<  0
              | cast<U32>(b | (a<<8)) << 16);
 #endif
 #endif
+#endif
 }
 
 LOWP_STAGE_PP(load_8888, const SkRasterPipelineContexts::MemoryCtx* ctx) {