On Sun, Jun 8, 2008 at 9:23 AM, Michael Fischer <[EMAIL PROTECTED]> wrote: > Hello Øyvind, > > do we really need the TRST signal to reset the > TAP controller, or can we reset the TAP by > software only?
TRST will also reset embedded ice registers and depending on the circuitry on the board other things could happen. We have a board where the JTAG circuitry is disconnected upon asserting TRST, so srst_only is then a lifesaver :-) You can use reset_config to disable the use of TRST for targets where this matter. This means that OpenOCD leaves this question to the target config file and it does not try to come up with one universal answer, or do I believe that one exists. > > If I take a look how the J-Link GDB is working, > I could not see a TRST low level. If we could > eliminate the need of the TRST signal, I think > we can remove the reset config too. Reset config tells OpenOCD if TRST was pulled by SRST, though it may be possible to detect this, although I don't know that it can be detected reliably. > > As I understand it correct, the TAP controller > can be reset by software too. > > Even we do not have the problem with the SRST and TRST > at the same time like: > https://lists.berlios.de/pipermail/openocd-development/2008-April/001720.html I added a rule to jtag_add_reset() that it could not change the state of srst and trst simultaneoulsy. Perhaps the API should be made more explicit by having separate jtag_add_trst(), jtag_add_srst() and jtag_add_tms() The complication is that srst pulls trst sometimes. I'd for one would love to get rid of reset_config if it could be done robustly. -- Øyvind Harboe http://www.zylin.com/zy1000.html ARM7 ARM9 XScale Cortex JTAG debugger and flash programmer _______________________________________________ Openocd-development mailing list [email protected] https://lists.berlios.de/mailman/listinfo/openocd-development
