On the TI OMAP3530 BSP, the following tests are known to fail with the issues described below:

 

Type of Test

Description

Audio Waveform Test (wavetest.dll) Subtest ID 3000 hangs

Networking - Ethernet tests (ndt_2c6.dll, perf_ndis6.dll, perf_winsock2.dll)

Subtest ID 5 (2c_StressSend) and subtest ID 6 (2c_StressReceive) in ndt_2c6.dll, subtest ID 1001 (Test 1) in perf_ndis6.dll, and a few tests in perf_winsock2.dll fail for the SMSC9118 chip Fast Ethernet Channel (FEC) on the TI OMAP3530. Restart the board after each test failure.

Verify DirectDraw test (DDrawTK.dll)

Subtest IDs 102 (Verify Surface Creation), 230,340 (Set/GetPixel Verification), 210,310 (ColorKey Blt), 200, 300 (Blt) fail on the TI OMAP3530. To work around this issue in subtest IDs 210, 310, 200, and 300 in ARM platforms with L2 cache, run the test in kernel mode by specifying the "-n" option in the command line. There is no workaround for the other failures.

DirectDraw Functional Test (DDFunc.dll) Subtests 304 and 404 failed as TI doesnt have Display Driver Configuration file which lists the supported surfaces on TI
DirectDraw Interface Tests (DDints.dll) Subtests 527 and 1795 fail

GDI tests (gdit.dll)

Subtest ID 222 (GradientFill) and subtest ID 231 (AlphaBlend) fail.







Several subtests fail if the test is run with following command line options - /Surface All or /Surface Extended

NLED Tests (nledtest.dll) Subtests 2102, 5013 and 6000-6012 fail







Subtests 5000, 5010, 7000, 7002, 7004, 7005, 7006 skip

USB function driver verification tests (usbfnbvt.dll)

Subtest ID 103 (Get/Set USB Function current client driver) and subtest ID 106 (Enumerate All Clients) fail.

SD card bus level functional tests (SDMemTux.dll)

Subtest IDs 1029 (Simple Synchronous Test of CMD24), 1030(Simple Asynchronous Test of CMD24), 1031 (Simple Synchronous Test of CMD25) and 1032 (Simple Asynchronous Test of CMD25) fail on the TI OMAP3530 device for MultiMediaCards (MMC).

File System Driver Test for SD Card (fsdtst.dll) If the test is run on SD card with 3 partitions, the subtests 5010, 5015 and 5016 fail.







However these subtests pass if the test is run on SD card with 1 or 2 partition

Power Manager Timeouts Test

Test ID 32 is known to fail on the TI_EVM3530 platform. The implementation of the RTC Alarm on the TI platform is handled through a separate SYSINTR which is shared by other hardware components.

Power Manager IOCTL Test     Subtests 1001, 1003 and 1004 fail

OALTestTimers (oaltesttimers.dll)

The following subtests fail:

  • Subtest ID 3010 (Compare All Three Timers - Busy Sleep)
  • Subtest ID 3020 (Compare All Three Timers - OS Sleep)
  • Subtest ID 3030 (Compare All Three Timers - Track OEMidle Periodic)
  • Subtest ID 3040 (Compare All Three Timers - Track OEMidle Random)
Serial Communication Tests (pserial.dll) In Master mode,







Following subtests fail -     1000, 2001







Following subtests abort -  14







Following subtests skip -     16-21, 23, 41-48, 1001-1003, 2000, 2002, 2003















In Slave Mode,







Following subtests fail -    41-48, 1000, 1001







Following subtests abort - 2001







Following subtests skip -    1002, 1003, 2000, 2002, 2003

USB loop back tests(usbtest.dll)







OTG port:

  • ISOCHRONOUS Transfer is not implemented on OTG Host driver. Because of this limitation few of the USB host tests will fail in full speed configuration for regular packet size.
  • Test cases to set packet size to small (TC# 3001) and non-regular  (TC# 3003) will hang.
  • With Lpbkcfg1 on CEPC, several subtests fail in High speed configuration
  • With Lpbkcfg2 on CEPC, subtests 1421 and 1441 fail

USB loop back tests(usbtest.dll)







EHCI port:

  • Due to a hardware limitation low/full speed devices cannot be connected to EHCI port. To workaround this issue, use a USB 2.0 hub between the USB port and the low/full speed device. Due to this limitation full speed tests are not applicable for this port. However full speed golden bridge tests can be run.
  • Test suite hangs at TC#9001 in high speed configuration for all packet sizes in both normal and golden bridge scenarios
  • Test cases to set packet sizes to small (TC# 3001) and non-regular (TC# 3003) packet sizes will hang in golden bridge set up for full speed configuration.
  • With Lpbkcfg1 on CEPC, several subtests fail in High speed configuration
  • With Lpbkcfg2 on CEPC, subtests 1421 and 1441 fail

USB performance test(usbperf.dll)







OTG port:

  • Subsets may fail in full speed configuration.

USB performance test(usbperf.dll)







EHCI port:

  • Test suite hangs at TC#1005 and it won’t proceed further in both normal and golden bridge scenario for high speed configuration.
  • Test suite hangs at TC#1001 and it won’t proceed further in full speed golden bridge scenarios.

On the TI OMAP3530 BSP, the following platform specific test (Non-CTK) is known to fail with the issues described below:

Type of Test

Description

Dynamic Voltage Frequency Scaling (DVFS) power management stress test (dvfs_stress_test)

Subtest 1 (DVFS Stress Table) in dvfs_stress_test causes the TI OMAP3530 device to stop responding.

 On the TI OMAP3530 BSP, the following are known issues:

Module

Issue Description

OpenGL Drivers

1) PowerVR drivers cannot be used with OpenGL compositor. Data abort exception is observed in nullws.dll on booting an image which has PowerVR drivers and OpenGL compositor.
To work around this problem use GDI compositor.
 
2) In systemidle state OpenGL driver stops rendering as expected but display does not turn off. Due to this, applications or tests which use OpenGL driver will appear to hang when device goes to SYSTEMIDLE state. On tapping the LCD screen, device comes back to ON power state and OpenGL driver resumes rendering.
 
3) Prefetch abort exception occurs when IE Browser uses OpenGL drivers for Hardware acceleration. Below is the scenario:

a)   Open IE browser

b)   Navigate to a web page for example, http://m.bing.com/

c)  Close IE browser
4) OpenGL driver does not synchronize eglswapbuffers with display vsync rate causing performance issues in SilverlightForWindowsEmbedded framework.

 


See Also