vgsp: MME setup prelude - capture client MME context-setup runs (0x114..0x124/0x234/0x550..0x568/0x3400..0x34fc) verbatim, exact-run dedup, 512-word buf + gen counter (sema_mme_prelude_cap); one-time-per-generation preplay in GR launch build (sema_gl_mme_prelude); capture verified - full macro program 11 SA/IR pairs ~206 code words, MME-PRELUDE emitted gen=5 words=444; no further mme_esr=80000004 after prelude (r1079); remaining: setup push unreadable=1 skip before pin demand; r1082-class content-independent dispatch stall - SESSION 153 r1078-r1082

1 files changed, 14 insertions, 0 deletions

diff --git a/kernel/gsp_sema_service.h b/kernel/gsp_sema_service.h
index 69d6f33..12a785c 100644
--- a/kernel/gsp_sema_service.h
+++ b/kernel/gsp_sema_service.h
@@ -448,6 +448,20 @@ struct gsp_sema_gr_reloc {
 int gsp_sema_gr_reloc_resolve(u64 va, struct gsp_sema_gr_reloc *out);
 
 /*
+ * SESSION 153-CONTINUED-53: MME context-setup prelude accessor for the GR
+ * launch builder. Returns the captured setup runs (C597 MME family
+ * 0x114..0x124/0x234/0x550..0x568/0x3400..0x34fc, verbatim, first-
+ * encounter order), the word count in *n and the capture generation in
+ * *gen. The builder preplays the words once per generation, ahead of the
+ * first 3D window replay, so the ladder context's MME has its macro
+ * program before CALL_MME_MACRO runs. Reading the generation before and
+ * after the copy detects a concurrent append (the service thread only
+ * appends; a mismatch just defers emission to the next launch). *n is 0
+ * until the client's setup push has been decoded.
+ */
+const u32 *gsp_sema_mme_prelude_get(u32 *n, u32 *gen);
+
+/*
  * r378 write guards. Both re-validate a guest-memory write target that was
  * resolved earlier (ledger/pin/walk/counter registration, or the GR submit's
  * per-page resolution) against the CURRENT resolution before the bytes land,