Updated 2/15/2014.
Ignore this information. I was wrong. The OpenSDA FW is just fine.
Long story short, turns out the problem was my music player application (Rhapsody) on Windows 7. Rhapsody was auto-detecting the connected OpenSDA FW device as a "media player / storage" device and doing something to it (writing to it?) that caused the OpenSDA FW to go into a state where the JTAG lines were incorrectly held low.
Not sure if there's a way to make the OpenSDA FW more robust to avoid a situation like this. Either way, my solution is to not use the locally installed Rhapsody application when doing active development with the Vybrid Tower HW.
================================================================================
================================================================================
I just got done resolving an issue wrt using Vybrid Tower board's (VF65GS10 rev G) J5 JTAG port that may help others. It was impacted by which OpenSDA FW was installed on the K20 part. (I just got done resolving an issue wrt using Vybrid Tower board's (VF65GS10 rev G) J5 JTAG port that may help others. It was impacted by which OpenSDA FW was installed on the K20 part. (forum thread)
Trying to use the (VF65GS10 rev G) J5 JTAG port with the OpenSDA CMSIS-DAP FW installed appears to be an undocumented, unsupported configuration. At this point, it appears it's best to keep the Virtual COM Port FW installed when using the J5 JTAG header to develop / debug SW for the Vybrid part.
The OpenSDA CMSIS-DAP FW was incorrectly pulling the Vybrid's JTAG TMS signal low instead of letting it float per JTAG interface requirements. This impacted the ability of the J-link to properly communicate with the Vybrid via JTAG. Once I restored the Virtual COM Port FW to the OpenSDA / K20 part, the TMS signal was restored to the properly pulled up, 3.3V level.
Looking at the Vybrid Tower board schematic (rev G), the SI_SPI0_SOUT signal coming out of the K20 part is tied to the JTMS signal (Vybrid pad PTA11). For some reason, the OpenSDA CMSIS-DAP FW is causing that signal to be driven low. I would have expected the SI_SPI0_SOUT signal to remain in a high impedance state when the CMSIS-DAP port was not being used to actively download / debug FW. Not sure if this is a bug in the CMSIS-DAP FW.
In summary, when using the J5 JTAG port to develop / debug, make sure OpenSDA Virtual COM Port FW is installed.