#25 Command 'reset' does not reset internal states

[Arnim Läuger]

When the 'reset' command is issued, a TAP reset of the chain is triggered by calling jtag_reset() but internal states of UrJTAG don't follow this update.
E.g. when the instruction EXTEST was selected for the currently active part, a 'print' still shows EXTEST although the TAP reset implicitly selects BYPASS.
This is probably true for other software states that aren't obvious.

[Arnim Läuger]

Logged In: YES
user_id=156321
Originator: YES

Should be fixed with SVN r972:
tap_reset() in tap.c sets instructions of all parts in the chain to BYPASS.

[Arnim Läuger]

Logged In: YES
user_id=156321
Originator: YES

The fix by r972 is not consistent with 1149 since the instruction that is established via TAP reset is IDCODE if implemented, and BYPASS only if IDCODE is not available.
This fix requires more intelligence, though.

[Arnim Läuger]

Logged In: YES
user_id=156321
Originator: YES

r1302: Established a new function tap_reset_bypass() performs a TAP reset and forces the BYPASS instruction into all parts of the chain (flooding all IRs with all-1). Documented by setting the active instruction of all parts to "BYPASS" (where defined).
Modules that require to perform a consistent TAP reset on a configured chain/cable should use this function instead of the bulk tap_reset() that simply resets all TAP controllers.