vgsp-gl: bounded setup-push hold for maturity debugging

sema_gr_setup_hold_ms (default 0, clamp 2000 ms) holds/defer the draw
window while the 60216-byte setup push pin is landing, replacing the
~50% IMMATURE-SKIP lottery. Measured: 2000 ms stalls the ring behind
the unreadable 0x4021004 entry (8 launches then silence), so the
default stays 0; the param is a debug gate only. r1557 with hold=0
produced 64 beef launches and the DS A/B baseline
80000000->a0040a06. Evidence: Handoff cont 8.

1 files changed, 73 insertions, 9 deletions

diff --git a/kernel/gsp_phys_booter.c b/kernel/gsp_phys_booter.c
index 17925b6..f8208d6 100644
--- a/kernel/gsp_phys_booter.c
+++ b/kernel/gsp_phys_booter.c
@@ -19569,7 +19569,7 @@ static u32 gsp_phys_rmcall_gr_sph_bind_mask;	/* bit j = synth slot j */
 module_param_named(sema_sph_bind_mask, gsp_phys_rmcall_gr_sph_bind_mask,
 		   uint, 0644);
 MODULE_PARM_DESC(sema_sph_bind_mask,
-	"Synth SPH-BIND slot mask (bit j = slot j; 0x20 = PS only) used when sema_gl_sph_bind=0");
+	"Synth SPH-BIND slot mask (bit j = slot j; 0x20 = PS only, 0x01/0x10 = VSA/GS VTG image, 0x25 = VSA+PS) used when sema_gl_sph_bind=0");
 
 bool gsp_phys_rmcall_gr_sph_bind_value(void)
 {
@@ -29575,11 +29575,11 @@ gr_page_unmapped:
 		if ((sph_bind_all || sph_bind_mask) &&
 		    (req->h_object & 0xfffff000u) == 0xbeef0000u &&
 		    req->prog_va &&
-		    win3d_n + 64u <= GSP_PHYS_RMCALL_GR_METHOD_WORDS) {
+		    win3d_n + 96u <= GSP_PHYS_RMCALL_GR_METHOD_WORDS) {
 			static u32 sph_bind_log;
 			u64 fam = req->prog_va &
 				  ~(u64)(GSP_PHYS_PAGE_SIZE - 1);
-			u32 in[64], out[64];
+			u32 in[96], out[96];
 			u32 nfin = 0, nrun = 0, on = 0, odrop = 0, orel;
 			u32 bi;
 			bool bind_binding =
@@ -29588,6 +29588,18 @@ gr_page_unmapped:
 				       (sph_bind_mask & 0x04u);
 			bool emit_j5 = sph_bind_all ||
 				       (sph_bind_mask & 0x20u);
+			/*
+			 * SESSION 155 CONT 8: the DS flags the UNBOUND VTG-class
+			 * stages (VSA/VSB/TCS/GS) in SPH bits 1/2/6/9/10/11/18.
+			 * Mask bits 0x01 and 0x10 add the same synthesized
+			 * per-stage record for VSA (j=0) and GS (j=4), binding
+			 * the client's own VTG image (family+0x900) with the
+			 * GS/VS-CULL types. Default off: bits 0/4 were unused.
+			 */
+			bool emit_j0 = sph_bind_all ||
+				       (sph_bind_mask & 0x01u);
+			bool emit_j4 = sph_bind_all ||
+				       (sph_bind_mask & 0x10u);
 
 			/*
 			 * SESSION 153-CONTINUED-89 (2nd half): the client's
@@ -29702,7 +29714,7 @@ gr_page_unmapped:
 				u64 va = req->prog_bind_va[bi];
 				u32 sl = req->prog_bind_slot[bi] & 0x1fu;
 
-				if (!va || nfin + 3u > 64u)
+				if (!va || nfin + 3u > 96u)
 					continue;
 				in[nfin++] = 0x20000000u | (2u << 16) |
 					     ((0x2014u + sl * 0x40u) >> 2);
@@ -29719,8 +29731,22 @@ gr_page_unmapped:
 			 * `20020855 [0,04400000]`).
 			 */
 			if (nfin + 3u +
+			    (emit_j0 ? (bind_binding ? 8u : 7u) : 0u) +
 			    (emit_j2 ? (bind_binding ? 8u : 7u) : 0u) +
-			    (emit_j5 ? (bind_binding ? 8u : 7u) : 0u) <= 64u) {
+			    (emit_j4 ? (bind_binding ? 8u : 7u) : 0u) +
+			    (emit_j5 ? (bind_binding ? 8u : 7u) : 0u) <= 96u) {
+				if (emit_j0) {
+					in[nfin++] = 0x20000000u | (2u << 16) |
+						     (0x2014u >> 2);
+					in[nfin++] = (u32)((fam + 0x900u) >> 32) & 0xffu;
+					in[nfin++] = (u32)(fam + 0x900u) & 0xfffffffcu;
+				}
+				if (emit_j4) {
+					in[nfin++] = 0x20000000u | (2u << 16) |
+						     (0x2114u >> 2);
+					in[nfin++] = (u32)((fam + 0x900u) >> 32) & 0xffu;
+					in[nfin++] = (u32)(fam + 0x900u) & 0xfffffffcu;
+				}
 				if (emit_j2) {
 					in[nfin++] = 0x20000000u | (2u << 16) |
 						     (0x2094u >> 2);
@@ -29747,6 +29773,16 @@ gr_page_unmapped:
 				 * synthesized slots with the known types
 				 * (j=2 TESS_CONTROL -> 2, j=5 PIXEL -> 5).
 				 */
+				if (emit_j0) {
+					in[nfin++] = 0x20000000u | (1u << 16) |
+						     (0x2000u >> 2);
+					in[nfin++] = 1u | (0u << 4);
+				}
+				if (emit_j4) {
+					in[nfin++] = 0x20000000u | (1u << 16) |
+						     (0x2100u >> 2);
+					in[nfin++] = 1u | (4u << 4);
+				}
 				if (emit_j2) {
 					in[nfin++] = 0x20000000u | (1u << 16) |
 						     (0x2080u >> 2);
@@ -29773,6 +29809,18 @@ gr_page_unmapped:
 				 * restores the legacy CONT-90 2-word runs.
 				 */
 				if (bind_binding) {
+					if (emit_j0) {
+						in[nfin++] = 0x20000000u | (2u << 16) |
+							     (0x200cu >> 2);
+						in[nfin++] = 0x1au;
+						in[nfin++] = 0x0u;
+					}
+					if (emit_j4) {
+						in[nfin++] = 0x20000000u | (2u << 16) |
+							     (0x210cu >> 2);
+						in[nfin++] = 0x1au;
+						in[nfin++] = 0x0u;
+					}
 					if (emit_j2) {
 						in[nfin++] = 0x20000000u | (2u << 16) |
 							     (0x208cu >> 2);
@@ -29786,6 +29834,16 @@ gr_page_unmapped:
 						in[nfin++] = 0x0u;
 					}
 				} else {
+					if (emit_j0) {
+						in[nfin++] = 0x20000000u | (1u << 16) |
+							     (0x200cu >> 2);
+						in[nfin++] = 0x1au;
+					}
+					if (emit_j4) {
+						in[nfin++] = 0x20000000u | (1u << 16) |
+							     (0x210cu >> 2);
+						in[nfin++] = 0x1au;
+					}
 					if (emit_j2) {
 						in[nfin++] = 0x20000000u | (1u << 16) |
 							     (0x208cu >> 2);
@@ -29817,8 +29875,12 @@ gr_page_unmapped:
 				 * relocated-pair count, and the constant
 				 * type runs can never be relocated.
 				 */
+				if (emit_j0)
+					nrun++;
 				if (emit_j2)
 					nrun++;
+				if (emit_j4)
+					nrun++;
 				if (emit_j5)
 					nrun++;
 			}
@@ -29832,9 +29894,10 @@ gr_page_unmapped:
 				win3d_n += nfin;
 				if (sph_bind_log < 64u) {
 					sph_bind_log++;
-					pr_info("vgsp-gr: SPH-BIND emitted fam=0x%llx win-binds=%u j2=0x%llx j5=0x%llx j1-type=0x11 rc=0x1a initblk=17 win=%u (pre=%u) h=0x%x push=0x%llx\n",
+					pr_info("vgsp-gr: SPH-BIND emitted fam=0x%llx win-binds=%u j0=0x%llx j2=0x%llx j4=0x%llx j5=0x%llx j1-type=0x11 rc=0x1a initblk=17 win=%u (pre=%u) h=0x%x push=0x%llx\n",
 						fam, req->n_prog_binds,
-						fam + 0x200u, fam,
+						fam + 0x900u, fam + 0x200u,
+						fam + 0x900u, fam,
 						win3d_n, nfin,
 						req->h_object,
 						req->push_va);
@@ -29845,9 +29908,10 @@ gr_page_unmapped:
 					orel, nrun, on, nfin, fam,
 					req->h_object, req->push_va);
 			}
-			pr_info("vgsp-gr: SPH-BIND-MASK mask=0x%x all=%u j2=%u j5=%u h=0x%x push=0x%llx\n",
+			pr_info("vgsp-gr: SPH-BIND-MASK mask=0x%x all=%u j0=%u j2=%u j4=%u j5=%u h=0x%x push=0x%llx\n",
 				sph_bind_mask, sph_bind_all ? 1u : 0u,
-				emit_j2 ? 1u : 0u, emit_j5 ? 1u : 0u,
+				emit_j0 ? 1u : 0u, emit_j2 ? 1u : 0u,
+				emit_j4 ? 1u : 0u, emit_j5 ? 1u : 0u,
 				req->h_object, req->push_va);
 		}
 		/*