On the Freescale i.MX27 BSP, the following CTK tests are known to fail with the issues described below:
Type of Test
Description
Communication bus tests - regular serial test (pserial.dll)
Subtest ID 1001 (Stress Test for 19200 bps) for COM port (UART) may stop working. This test case is a stress test at baud rate 19,200 and runs for 100 iterations.
Serial communication test over USB Serial function aborts in synchronization in master and slave mode. As a workaround run the test in kernel mode. In kernel mode Subtest ID 20 (TestXonXoffIdle ) and subtest ID 21 (TestXonXoffReliable ) may stop working for USB serial function in subordinate mode if the subordinate is not able to receive data at the same speed the master is sending.
Performance subtests (IDs 2000-2003) fail for USB serial host in the Freescale i.MX27 in subordinate mode due to a failure in DATA integrity check.
Winsock performance tests (perf_winsock2.dll)
Winsock performance tests fail for RNDISFN.
RNDISFN tests
Ping between RNDISFN and RNDIS Host fails when done with a large packet size. The ping works with the default packet size of 32 bytes, but if you increase the packet size to 10000, the ping may fail.
Two-card network card miniport driver (NDIS 6) test (ndt_2c6.dll) and Winsock performance test (perf_winsock2.dll)
The ndt_2c6 subtest ID 1 (2c_sendpackets) and perf_winsock2 test give "Babling TX" errors and then fail for platforms with the Fast Ethernet Channel (FEC) driver enabled.
Audio tests - audio playback and record verification (wavetest.dll)
Subtest ID 3008 (Capture Multiple Streams) fails inconsistently when run several times.
Audio tests -audio playback verification tests and audio playback and record verification tests (wavetest.dll)
Audio wavetest subtest ID 6000(playback mixing) and subtest ID 6001 (capture mixing) fail inconsistently while playing the media in the expected time.
Audio tests - Audio drift test (drifttest.dll)
Audio drift test (drifttest.dll) fails due to the difference in algorithms used in resampling and the basic hardware limitation.
OAL - IOCTL tests (oaltestioctls.dll)
Subtest ID 5502 (test that IOCTL_HAL_GET_WAKE_SOURCE buffers fail at appropriate times) fails with the error "Wake Source is not valid"; the wake source returns 255.
OAL - RTC Alarm test (oaltestrtcalarm.dll)
Subtest ID 1003 (Verify RTC alarm fires at user provided time and duration) fails.
Multimedia - camera tests (CameraGraphTests.dll)
Subtest ID 401(Still image verification tests, still image without preview) fails.
Subtest ID 1000 (VideoControl GetCaps test) fails.
Camera IO test (camiotests.dll)
Subtest ID 1000 (Test by creating a still pin) fails. The camera driver only supports two physical pins, the PREVIEW and CAPTURE pins. The STILL pin borrows its signal from the PREVIEW pin. If the PREVIEW pin is not running, the STILL pin will not work. In this test, the PREVIEW pin is not started and the STILL pin test fails as a result.
Backlight API test (backlightapitest.dll)
Subtest ID 1006 (Power Manager Backlight Operations) is expected to skip
Rotation Scaling VMR test (rotationscalingtest.dll)
The following subtests are expected to fail:
Power Timeout tests (pwrtimeouts.dll)
Power States tests (pwrtststates.dll)
Video capture test (vidcaptest)
In the video capture test (vidcaptest), subtest 14 (PinFlowingTests) fails.
USB Test (usbtest.dll)
The following subtests fail in full speed mode because the i.MX27 platform uses a cloned EHCI driver not based on the Windows Embedded Compact 7 common driver. These subtests set the packet size before running further USB tests. Although the subtests fail, the packet size is set properly.
Several subtests fail for non-regular packet size in full-speed mode and High-speed mode.
Subtest ID 9001 fails resulting the subtest IDs 9003-9991 to skip in both high-speed and full-speed mode. To work around this issue, run the usbtest.dll suite by excluding Subtest ID 9001. Subtest ID 9003-9007 passes while Subtest ID 9991 fails.
Several subtests are failing in golden bridge test in Full-speed mode.
On the Freescale i.MX27 BSP, the following platform specific tests (Non-CTK) are known to fail with the issues
described below:
Enhanced multimedia accelerator (eMMA) post -processor test (pptest.dll)
The eMMA post-processor test (pptest.dll) fails to convert YUV420 to RGB888 while resizing the width in the range of 1:1 to 1: 1.64.
Maheshkumar S Tiwari edited Revision 11. Comment: Added tags
Ed Price MSFT edited Revision 1. Comment: I retitled this. It seems to be about CTK test cases. Please review the title I used and adjust it if this doesn't make sense.
singwee edited Revision 5. Comment: Minor change to USB section.
Ishu Agarwal edited Revision 6. Comment: Updated the issues in USBtest.dll and usbperf.dll.
Ishu Agarwal edited Revision 8. Comment: Updated the description in pserial over USB function.
Siddhartha1 edited Revision 9. Comment: Added the failures in GDIT