Snapshot: fog implementation + fidelity tooling baseline (pre bilinear-clamp fix)

Per-vertex GS fog end-to-end (gs_stub emit incl. persp_emit5, gs_prim_list_feeder
XYZ2->XYZF2 on PRIM.FGE, gs_make_sh3_scheduler_fixture.py F/FGE packing), new fog
TBs, fidelity attribution tooling. Functional baseline before removing the dead
bilinear lerp8 clamps (Codex: 161-node comb loop -> -0.042ns setup fail).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

rtl/gif_gs/gs_persp_uv.sv

@@ -48,7 +48,9 @@ module gs_persp_uv #(
     input  logic [Q_W-1:0]     q,
     output logic               out_valid,
     output logic [TEXEL_W-1:0] u,
-    output logic [TEXEL_W-1:0] v
+    output logic [TEXEL_W-1:0] v,
+    output logic [3:0]         u_frac,
+    output logic [3:0]         v_frac
 );
     localparam int RLAT = 3;  // gs_reciprocal_stub latency
 
@@ -96,13 +98,23 @@ module gs_persp_uv #(
             out_valid <= 1'b0;
             u         <= '0;
             v         <= '0;
+            u_frac    <= '0;
+            v_frac    <= '0;
         end else begin
             logic [PROD_W-1:0] u_prod, v_prod;
+            logic [PROD_W-1:0] u_fixed4, v_fixed4;
             out_valid <= recip_valid;
             u_prod    = uq_pipe[RLAT-1] * w_recip;
             v_prod    = vq_pipe[RLAT-1] * w_recip;
             u         <= clamp_texel(u_prod);
             v         <= clamp_texel(v_prod);
+            // Preserve four fractional texel bits for the palette-bilinear
+            // path.  The legacy integer outputs remain the exact SCALE-bit
+            // truncation above; these are simply the next four product bits.
+            u_fixed4 = u_prod >> (SCALE - 4);
+            v_fixed4 = v_prod >> (SCALE - 4);
+            u_frac <= u_fixed4[3:0];
+            v_frac <= v_fixed4[3:0];
         end
     end