https://bugs.winehq.org/show_bug.cgi?id=52164
Bug ID: 52164 Summary: mfplat:mfplat gets stuck in test_MFCreateDXSurfaceBuffer() on cw-gtx560 (Nvidia) Product: Wine Version: unspecified Hardware: x86-64 OS: Windows Status: NEW Severity: normal Priority: P2 Component: mfplat Assignee: wine-bugs@winehq.org Reporter: fgouget@codeweavers.com
mfplat:mfplat gets stuck on cw-gtx560 (Nvidia):
07b4:mfplat: 6 tests executed (0 marked as todo, 0 failures), 0 skipped. 1760:mfplat: 8 tests executed (0 marked as todo, 0 failures), 0 skipped. 1ab4:mfplat: 8 tests executed (0 marked as todo, 0 failures), 0 skipped. 0668:mfplat: 8 tests executed (0 marked as todo, 0 failures), 0 skipped. mfplat:mfplat:1ae8 done (258) in 120s
(258 == WAIT_TIMEOUT) https://test.winehq.org/data/patterns.html#mfplat:mfplat
Running with WINETEST_REPORT_SUCCESS shows that the freeze happens in test_MFCreateDXSurfaceBuffer(), specifically in the first IMFMediaBuffer_Unlock() call.
hr = IMFMediaBuffer_Lock(buffer, &data, NULL, &length); ok(hr == S_OK, "Unexpected hr %#x.\n", hr); ok(length == max_length, "Unexpected length.\n"); // <--- last trace
/* Unlock twice. */ hr = IMFMediaBuffer_Unlock(buffer); ok(hr == S_OK, "Unexpected hr %#x.\n", hr); // <--- never seen
This is specific to cw-gtx560, the TestBot's only Nvidia test configuration. It also only happens on Windows 10, from 1709 to 21H1: the test succeeds on cw-gtx560 when running Windows 8.1 despite the fact that it still runs test_MFCreateDXSurfaceBuffer() (the skip trace is not present in the win81 WineTest report).