sema-svc: apply releases promptly and decode compute-push releases

The content sweep now:
- decodes a release whose method stream continues with any INC_METHOD,
  not only the CE LAUNCH_DMA group, so the releases at the head of a
  compute/GR channel's first push are applied too;
- remembers releases whose target VA had no translation (deferred list)
  and retries them every service pass, so a VA->PA injection lands them
  immediately instead of after the next full sweep;
- re-arms the fast sliced sweep for a few seconds after an injection;
- treats a channel whose USERD/ring cannot be resolved as stalled, so the
  sweep keeps looking for the releases its unreadable pushes carry.

tools: add per-cdt client-pin helper (sem pages + /dev/nvidia0 windows),
sem-page scanner and memprobe fixes (short-read guard, --brief).
Live-verified: injecting a client sem page applied 12 deferred releases
within a second; the service kept applying (89 -> 202) while cdt ran.

1 files changed, 8 insertions, 3 deletions

diff --git a/tools/guest-diag/memprobe.py b/tools/guest-diag/memprobe.py
index 26ed0af..96dd40b 100644
--- a/tools/guest-diag/memprobe.py
+++ b/tools/guest-diag/memprobe.py
@@ -126,8 +126,10 @@ def dump_tracker(pid, fd, p):
     if sob:
         print("  sem_obj[0:0x40]=%s" % sob.hex())
         mem = r64(sob, 0x10)
-        mvb = rd(fd, mem & ~0xfff, 0x40)
-        print("  sem_mem=0x%x val@+0x%x=0x%x" % (mem, mem & 0xfff, r32(mvb, mem & 0xfff) if mvb else -1))
+        mvb = rd(fd, mem & ~0xfff, 0x1000)
+        off = mem & 0xfff
+        val = r32(mvb, off) if mvb and off + 4 <= len(mvb) else -1
+        print("  sem_mem=0x%x val@+0x%x=0x%x" % (mem, off, val))
     q = r64(rd(fd, p + 0x48, 8) or b"\0" * 8)
     print("  prog_obj=0x%x next=0x%x cblist=0x%x idx95bc=0x%x p34=0x%x"
           % (q, r64(rd(fd, p + 0x9470, 8) or b"\0" * 8),
@@ -154,15 +156,18 @@ def main():
         if mode == "scan":
             lo, hi = 0x204800000, 0x204c0000
             limit = 8
+            brief = False
             for i, a in enumerate(sys.argv[3:]):
                 if a == "--min": lo = int(sys.argv[4 + i], 16)
                 if a == "--max": hi = int(sys.argv[4 + i], 16)
                 if a == "--limit": limit = int(sys.argv[4 + i])
+                if a == "--brief": brief = True
             res = find_trackers(pid, fd, lo, hi, limit)
             for p, sub, last, cur, sem, mem in res:
                 print("P=0x%x sub=0x%x last=0x%x cur=0x%x sem_obj=0x%x sem_mem=0x%x" %
                       (p, sub, last, cur, sem, mem or 0))
-                dump_tracker(pid, fd, p)
+                if not brief:
+                    dump_tracker(pid, fd, p)
         elif mode == "dump":
             dump_tracker(pid, fd, int(sys.argv[3], 16))
         elif mode == "hex":