tools: pin GR push-buffer window from guest pagemap per boot; widen injected-pin table

The GR channel GPFIFO entries point at pushes in the client GPU VA window
(observed 0x200400000 + stride, one allocation per channel) and the pump
cannot read them: pushes_nomap grows and the tracking-semaphore releases
they carry never land. The pushes are sysmem pages mapped into the client
process, so /proc/<pid>/pagemap in the guest gives the GPA.

New guest helper guest_pmap_win.py batch-reads the pagemap of a VA range
(512 entries per pread). New host tool pin_gr_window.py rescans the window
on an interval, follows cdt pid changes, and injects VA->PA runs through
sema_map_add (same mechanism as the counters and the CE pushbuffer pins).
The injected-pin table grows 8k -> 32k slots so the CE pushbuffer (4096)
plus the GR window pages fit without drops.

1 files changed, 5 insertions, 2 deletions

diff --git a/kernel/gsp_sema_service.c b/kernel/gsp_sema_service.c
index b9e7db3..1296f81 100644
--- a/kernel/gsp_sema_service.c
+++ b/kernel/gsp_sema_service.c
@@ -178,9 +178,12 @@ static u32 svc_map_hash[SEMA_MAP_MAX * 2];
 #define SVC_PIN_PROBE		128
 /*
  * Host-injected translations (debug `sema_map_add`): authoritative, never
- * overwritten by self-discovery, checked before every other source.
+ * overwritten by self-discovery, checked before every other source. The GR
+ * push-buffer window pins thousands of client pages per boot (one entry per
+ * 4 KB page), so the table must hold them plus the CE push-buffer region
+ * (4096 pages) without dropping inserts.
  */
-#define SVC_IPIN_SLOTS		(1u << 13)
+#define SVC_IPIN_SLOTS		(1u << 15)
 #define SVC_IPIN_PROBE		64
 
 struct svc_pin {