[PATCH 0/1] MR5368: win32u: Avoid a crash when cleaning up a monitor without an adapter.

[Rémi Bernon (＠rbernon)]

I *think* this will fix the test timeouts that happen on Gitlab since https://gitlab.winehq.org/wine/wine/-/merge_requests/5262.

I'm not completely sure that it will, because the failure is very elusive and somehow only happening on Gitlab CI, but there's a logic behind it: we were previously enumerating monitors from the adapters we found, so every monitor enumerated had a matching adapter. We are now enumerating monitors directly, and matching their adapter next.

How we then end up with a monitor without a matching adapter is still unclear, but I think it's better to fix it like this for now, as it is plaguing MRs with failures and long timeouts.

[PATCH 1/1] win32u: Avoid a crash when cleaning up a monitor without an adapter.

[Rémi Bernon]

From: Rémi Bernon

---
 dlls/win32u/sysparams.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/dlls/win32u/sysparams.c b/dlls/win32u/sysparams.c
index 07b9d46d9eb..d5ad97a2c33 100644
--- a/dlls/win32u/sysparams.c
+++ b/dlls/win32u/sysparams.c
@@ -1501,7 +1501,7 @@ static void clear_display_devices(void)
     while (!list_empty( &monitors ))
     {
         monitor = LIST_ENTRY( list_head( &monitors ), struct monitor, entry );
-        adapter_release( monitor->adapter );
+        if (monitor->adapter) adapter_release( monitor->adapter );
         list_remove( &monitor->entry );
         free( monitor );
     }