sema-svc: persist GR counter slot per channel; stable slot layout; bulk pin queue 4k

- counter_pa autoreg result is written back to the channel table so later
  passes cannot re-target a different slot when the live GR count changes
  during teardown; re-registration of the same channel preserves the slot.
- counter slot layout uses the max GR count ever seen in the generation
  instead of the shrinking live count.
- sema_map_add queue 64 -> 4096 entries so a full-window pin burst drains
  without ENOSPC retries.
- pin_gr_stream: one batched sudo pass per flush (was one process per run,
  ~4 s for 400 runs), retrying only queue-rejected runs.
- semfix_uvmt: SEMFIX_SKIP_PB=1 skips the blocking push-buffer chain walk
  when the vsock window stream already supplies those pages.

1 files changed, 7 insertions, 0 deletions

diff --git a/tools/guest-diag/semfix_uvmt.py b/tools/guest-diag/semfix_uvmt.py
index 02d4f9e..fd58299 100755
--- a/tools/guest-diag/semfix_uvmt.py
+++ b/tools/guest-diag/semfix_uvmt.py
@@ -43,6 +43,11 @@ COUNTER_LO = int(os.environ.get("COUNTER_LO", "0x20400f000"), 16)
 COUNTER_HI = int(os.environ.get("COUNTER_HI", "0x204010000"), 16)
 SEMSTREAM = os.environ.get("SEM_STREAM_FILE", "/tmp/opencode/sem_stream.txt")
 TRACE_EVERY = int(os.environ.get("UVMT_TRACE_EVERY", "5"))
+# The push-buffer resolution walks the channel/pool/manager chain with many
+# QMP reads and can block the poller for tens of seconds. When the GR window
+# stream (pin_gr_stream.py) already supplies the push-buffer pages, skip it so
+# the tracking-sem poller stays live during the client's deadline window.
+SKIP_PB = os.environ.get("SEMFIX_SKIP_PB", "0") == "1"
 
 
 class Stream:
@@ -309,6 +314,8 @@ def main():
                         ok = pin(gv, gpa)
                         log("semfix: sem=0x%x pool va=0x%x -> gpa=0x%x pin=%s"
                             % (sem, gv, gpa, "ok" if ok else "FAILED"))
+                    if SKIP_PB:
+                        continue
                     d = {}
                     try:
                         res = resolve_from_sem(q, sem, d)