vgsp: stop fabricating 0xFFFF into the nvPush error notifier; add r1780 GL ops tooling

The uncommitted keep-alive change wrote 0xFFFF into slot 0's status field
(u16 at notifier+0x0e) on the X DDX / GLX kick-idle channel families. Ground
truth (570.144 reference sources + stock nvidia_drv.so disassembly): slot 0
of an nvPush notifier block is the ERROR notifier; nvPushCheckChannelError
reads that status and 0xFFFF aborts every nvPushWaitForNotifier/idle wait as
a channel error. There is no pb_kickIfIdle in 570.144 - the DDX DMA-idle
test reads the progress-tracker semaphore word, not this slot. Leave slot 0
untouched; keep the 0x8000->0 completion ack for the remaining slots and the
r864 0xFFFF/0x8000 completion on the RC-watchdog/beef families. Correct the
CONT.408/409 comments accordingly.

New host/guest tooling: gpa.py (QMP gva2gpa), tick_gen.py (guest-RM timer
BO tick), guest_mode.sh (shared nvClip modeNum reader), gl_verify.sh
(X/GLX bring-up with ScreenInit-lottery retries), push_gl_tools.sh, the
gzdoom_pbuf_run.sh guest recipe and the test_tri SDL2 harness validation
program; ignore tools/guest-diag/build/.

1 files changed, 74 insertions, 0 deletions

diff --git a/tools/guest-diag/push_gl_tools.sh b/tools/guest-diag/push_gl_tools.sh
new file mode 100755
index 0000000..4064d58
--- /dev/null
+++ b/tools/guest-diag/push_gl_tools.sh
@@ -0,0 +1,74 @@
+#!/bin/bash
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright (c) 2026 Open-IOV contributors
+# push_gl_tools.sh - push the guest-side GL/GZDoom harness binaries and scripts
+# into the running guest's /tmp and chmod +x them.  Uses the established QGA
+# guest-exec + base64 mechanism from push_guest_tools.sh (scripts/vgpu_vm.py).
+#
+# Usage:
+#   bash tools/guest-diag/push_gl_tools.sh              # fixed list below
+#   bash tools/guest-diag/push_gl_tools.sh FILE...      # arbitrary host files
+#   bash tools/guest-diag/push_gl_tools.sh --dry-run    # list only, no push
+#
+# Destination is /tmp/<basename>.  The guest's /tmp is wiped on boot; run this
+# after every VM start.  It does NOT run the pushed programs.
+set -u
+
+HERE="$(cd "$(dirname "$0")" && pwd)"
+VMEXEC="$HERE/../../scripts/vgpu_vm.py"
+BUILD="$HERE/build"
+DEST_DIR="${DEST_DIR:-/tmp}"
+
+DRY_RUN=0
+if [ "${1:-}" = "--dry-run" ]; then
+    DRY_RUN=1
+    shift
+fi
+
+FIXED=(
+    "$BUILD/gl_frames"
+    "$BUILD/gz_pbuf.so"
+    "$BUILD/sdl_cap.so"
+    "$BUILD/test_tri"
+    "$HERE/guest_mode.sh"
+    "$HERE/gzf2_video.py"
+    "$HERE/gzdoom_pbuf_run.sh"
+)
+
+if [ "$#" -gt 0 ]; then
+    FILES=("$@")
+else
+    FILES=("${FIXED[@]}")
+fi
+
+if [ ! -x "$VMEXEC" ] && [ ! -f "$VMEXEC" ]; then
+    echo "push_gl_tools: vgpu_vm.py not found: $VMEXEC" >&2
+    exit 2
+fi
+
+push_b64() {   # push_b64 <local file> <remote file>
+    local src="$1" dst="$2" b64
+    b64=$(base64 -w0 "$src")
+    if [ "$DRY_RUN" = "1" ]; then
+        echo "PUSH-DRY $src -> $dst ($(stat -c%s "$src") bytes)"
+        return 0
+    fi
+    if ! sudo "$VMEXEC" exec --timeout 120 \
+        "echo '$b64' | base64 -d > '$dst' && chmod +x '$dst'" \
+        >/dev/null 2>&1; then
+        echo "PUSH-FAIL $src -> $dst" >&2
+        return 1
+    fi
+    echo "PUSHED $src -> $dst"
+}
+
+rc=0
+for f in "${FILES[@]}"; do
+    if [ ! -f "$f" ]; then
+        echo "PUSH-SKIP missing $f" >&2
+        rc=1
+        continue
+    fi
+    push_b64 "$f" "$DEST_DIR/$(basename "$f")" || rc=1
+done
+exit "$rc"