hil: silence false vid_pid warning for jlink args, tighten recover lintshil-openocd-recover-flasher

1 files changed, 9 insertions, 2 deletions

diff --git a/test/hil/hil_flash.py b/test/hil/hil_flash.py
index 2ff6b8b64..30850300b 100755
--- a/test/hil/hil_flash.py
+++ b/test/hil/hil_flash.py
@@ -100,10 +100,14 @@ def _openocd_cmd_base(flasher):
                 print(f'warning: {uid} has a malformed vid_pid {flasher["vid_pid"]!r} '
                       f'(want "0xVVVV 0xPPPP"); probe pin DROPPED, so discovery will open '
                       f'foreign usbfs nodes', file=sys.stderr, flush=True)
-    elif flasher.get('uid') not in _VID_PID_WARNED:
+    elif ('interface/jlink.cfg' not in (flasher.get('args') or '')
+          and flasher.get('uid') not in _VID_PID_WARNED):
         # stderr, once per probe: test_example captures stdout, so a passing run would
         # swallow this and the operator would never learn discovery still opens every
-        # usbfs node
+        # usbfs node. Skipped over the jlink driver: libjaylink discovery gates on
+        # idVendor == 0x1366 before libusb_open (same rationale as convoy_safe's
+        # docstring), so it never opens a foreign node regardless of the missing pin --
+        # the warning would be false there.
         _VID_PID_WARNED.add(flasher.get('uid'))
         print(f'warning: openocd flasher {flasher.get("uid", "?")} has no vid_pid pin; '
               f'probe discovery will open every usbfs node (hangs on a wedged one)',
@@ -149,6 +153,9 @@ def flash_openocd_seq(board, firmware, timeout=None):
 
 
 def reset_openocd_seq(board, timeout=None):
+    # Behaviorally reset_openocd, but this exact invocation is what was bench-validated
+    # over the jlink transport on all seven flasher_recover boards — don't fold into
+    # reset_openocd (or change either) without re-benching the recovery path.
     flasher = board['flasher']
     return hil_util.run_cmd(
         f'{_openocd_cmd_base(flasher)} -c "init" -c "reset run" -c "shutdown"',