riscv64/mc: Add VLEN=256 8bpc RVV blend_v function

SpacemiT K1

blend_v_w2_8bpc_c:       217.0 ( 1.00x)
blend_v_w2_8bpc_rvv:     143.3 ( 1.51x)
blend_v_w4_8bpc_c:       921.6 ( 1.00x)
blend_v_w4_8bpc_rvv:     236.3 ( 3.90x)
blend_v_w8_8bpc_c:      1738.2 ( 1.00x)
blend_v_w8_8bpc_rvv:     238.1 ( 7.30x)
blend_v_w16_8bpc_c:     3376.1 ( 1.00x)
blend_v_w16_8bpc_rvv:    298.0 (11.33x)
blend_v_w32_8bpc_c:     6648.0 ( 1.00x)
blend_v_w32_8bpc_rvv:    409.5 (16.24x)

src/riscv/64/mc.S

@@ -209,6 +209,27 @@ L(blend_h_epilog):
   jr t5
 endfunc
 
+function blend_v_vl256_8bpc_rvv, export=1, ext=v
+  li t0, 4
+  ble a3, t0, 24f
+  li t0, 8
+  beq a3, t0, 8f
+  li t0, 16
+  beq a3, t0, 16f
+32:
+  vsetivli zero, 24, e8, m1, ta, ma
+  j L(blend_v_epilog)
+16:
+  vsetivli zero, 12, e8, mf2, ta, ma
+  j L(blend_v_epilog)
+8:
+  vsetivli zero, 6, e8, mf4, ta, ma
+  j L(blend_v_epilog)
+24:
+  vsetvli zero, a3, e8, mf8, ta, ma
+  j L(blend_v_epilog)
+endfunc
+
 function blend_v_8bpc_rvv, export=1, ext=v
   li t0, 2
   beq a3, t0, 2f
@@ -220,19 +241,19 @@ function blend_v_8bpc_rvv, export=1, ext=v
   beq a3, t0, 16f
 32:
   vsetivli zero, 24, e8, m2, ta, ma
-  j 0f
+  j L(blend_v_epilog)
 16:
   vsetivli zero, 12, e8, m1, ta, ma
-  j 0f
+  j L(blend_v_epilog)
 8:
   vsetivli zero, 6, e8, mf2, ta, ma
-  j 0f
+  j L(blend_v_epilog)
 4:
   vsetivli zero, 3, e8, mf4, ta, ma
-  j 0f
+  j L(blend_v_epilog)
 2:
   vsetivli zero, 1, e8, mf8, ta, ma
-0:
+L(blend_v_epilog):
   csrw vxrm, zero
   la t1, dav1d_obmc_masks
   add t1, t1, a3

src/riscv/mc.h

@@ -34,6 +34,7 @@ decl_blend_dir_fn(BF(dav1d_blend_v, rvv));
 
 decl_blend_fn(BF(dav1d_blend_vl256, rvv));
 decl_blend_dir_fn(BF(dav1d_blend_h_vl256, rvv));
+decl_blend_dir_fn(BF(dav1d_blend_v_vl256, rvv));
 
 static ALWAYS_INLINE void mc_dsp_init_riscv(Dav1dMCDSPContext *const c) {
   const unsigned flags = dav1d_get_cpu_flags();
@@ -48,6 +49,7 @@ static ALWAYS_INLINE void mc_dsp_init_riscv(Dav1dMCDSPContext *const c) {
   if (dav1d_get_vlen() >= 256) {
     c->blend = BF(dav1d_blend_vl256, rvv);
     c->blend_h = BF(dav1d_blend_h_vl256, rvv);
+    c->blend_v = BF(dav1d_blend_v_vl256, rvv);
   }
 #endif
 }