Testsuite: T9214: make GRUB console-select navigation non-fatal on auto-boot race

Commit 7c33698 ("Testsuite: T9214: fix GRUB auto-boot race in console-select
navigation") narrowed the race between GRUB's own auto-boot countdown and this
script's post-install "Boot options" navigation, but did not close it: the
countdown is timed from when GRUB itself draws the menu, not from when this
script's regex match on the menu banner returns, so under host load GRUB can
still auto-boot before we react. When that happens, the subsequent
child.expect('Select console type', ...) blocks for the full timeout waiting
for a submenu that was never entered, then raises and aborts the test even
though the default entry already booted a working serial console (the
installer always answers 'S' to the console-type prompt).

Wrap the submenu navigation in a try/except so losing this race just logs a
warning and falls through instead of crashing, letting the already-auto-booted
default entry carry on into the login wait that already tolerates this case.

Author: c-po

scripts/check-qemu-install

@@ -547,7 +547,7 @@ def toggleUEFISecureBoot(c):
     UEFIKeyPress(c, KEY_DOWN)
     UEFIKeyPress(c, KEY_RETURN)
 
-def BOOTLOADERchooseSerialConsole(child, live: bool) -> None:
+def BOOTLOADERchooseSerialConsole(child, live: bool, log=None) -> None:
     """ Select GRUB boot entry that uses the serial console. This differs
         between a LIVE ISO image and an already installed system. """
     BOOTLOADER_TMO = 40
@@ -596,39 +596,58 @@ def BOOTLOADERchooseSerialConsole(child, live: bool) -> None:
         # Wait for GRUB
         child.expect(GRUB_STRING, timeout=BOOTLOADER_TMO)
 
-        # Unlike the live ISO menus (10s timeout), the installed system's
-        # top-level menu auto-boots its default entry after ~BOOTLOADER_LOAD_TMO
-        # seconds - the same delay this script would otherwise sleep before
-        # sending any key. Under host load that leaves no margin: GRUB can
-        # auto-boot before "Boot options" is ever selected. Send the first
-        # navigation key immediately to cancel the countdown, then give the
-        # screen time to settle.
-        # Select GRUB serial console
-        # Boot options
-        child.send(KEY_DOWN)
-        time.sleep(BOOTLOADER_LOAD_TMO)
-        child.send(KEY_RETURN)
-        time.sleep(BOOTLOADER_SLEEP)
-        # GRUB submenus never time out on their own, so confirm we actually
-        # landed on this submenu before navigating further - otherwise a
-        # dropped keypress leaves the VM stuck here until the login wait
-        # elsewhere expires
-        child.expect('Select console type', timeout=BOOTLOADER_TMO)
-
-        # Select console type
-        child.send(KEY_DOWN)
-        time.sleep(BOOTLOADER_SLEEP)
-        child.send(KEY_RETURN)
-        time.sleep(BOOTLOADER_SLEEP)
-        child.expect(r'ttyS \(serial\)', timeout=BOOTLOADER_TMO)
-
-        # *ttyS (serial)
-        child.send(KEY_DOWN)
-        time.sleep(BOOTLOADER_SLEEP)
-        child.send(KEY_RETURN)
-        time.sleep(BOOTLOADER_SLEEP)
-        # Boot
-        child.send(KEY_RETURN)
+        # The installed system's top-level menu auto-boots its default
+        # entry after ~BOOTLOADER_LOAD_TMO seconds, timed from when GRUB
+        # itself draws the menu - not from when this script's regex match
+        # on GRUB_STRING returns. Under host load the menu text can reach
+        # us well after that internal countdown already started, so there
+        # is no reliable amount of "send a key fast enough" that wins this
+        # race every time.
+        #
+        # That's fine to lose: the installer always answers 'S' (serial)
+        # to "What console should be used by default?", so the default
+        # entry GRUB auto-boots already targets the right console.
+        # waitForLogin()/loginVM(), called after this function returns,
+        # already tolerate landing straight on the GRUB countdown or the
+        # login prompt. So treat this submenu navigation as best-effort:
+        # if we don't land in "Boot options" in time, stop navigating and
+        # let the default entry (which is already auto-booting) carry on,
+        # instead of raising and aborting the whole test.
+        try:
+            # Select GRUB serial console
+            # Boot options
+            child.send(KEY_DOWN)
+            time.sleep(BOOTLOADER_LOAD_TMO)
+            child.send(KEY_RETURN)
+            time.sleep(BOOTLOADER_SLEEP)
+            # GRUB submenus never time out on their own, so confirm we actually
+            # landed on this submenu before navigating further - otherwise a
+            # dropped keypress leaves the VM stuck here until the login wait
+            # elsewhere expires
+            child.expect('Select console type', timeout=BOOTLOADER_TMO)
+        except pexpect.TIMEOUT:
+            if log is not None:
+                log.warning('GRUB auto-booted the default entry before "Boot '
+                            'options" navigation completed; continuing since '
+                            'the default entry already boots the serial '
+                            'console selected during install')
+        else:
+            # We're inside the submenu, so a timeout past this point is a
+            # real navigation bug, not the auto-boot race - let it propagate.
+            # Select console type
+            child.send(KEY_DOWN)
+            time.sleep(BOOTLOADER_SLEEP)
+            child.send(KEY_RETURN)
+            time.sleep(BOOTLOADER_SLEEP)
+            child.expect(r'ttyS \(serial\)', timeout=BOOTLOADER_TMO)
+
+            # *ttyS (serial)
+            child.send(KEY_DOWN)
+            time.sleep(BOOTLOADER_SLEEP)
+            child.send(KEY_RETURN)
+            time.sleep(BOOTLOADER_SLEEP)
+            # Boot
+            child.send(KEY_RETURN)
 
     return None
 
@@ -904,7 +923,7 @@ try:
         log.info('Disable UEFI Secure Boot for initial installation')
         toggleUEFISecureBoot(c)
 
-    BOOTLOADERchooseSerialConsole(c, live=(not args.cloud_init))
+    BOOTLOADERchooseSerialConsole(c, live=(not args.cloud_init), log=log)
     loginVM(c, log)
 
     #################################################
@@ -1070,7 +1089,7 @@ try:
     # Booting installed system
     #################################################
     log.info('Booting installed system')
-    BOOTLOADERchooseSerialConsole(c, live=False)
+    BOOTLOADERchooseSerialConsole(c, live=False, log=log)
 
     #################################################
     # Logging into VyOS system