Re: [Openocd-development] [PATCH] process reset - ready for test

[Ø. H. <oyv...@zy...>]

The general direction looks good to me.

I've made some commits and here are some commits about the patch:

- remove jtag cmd queue command from tcl interface. ref earlier
discussions it is
the job of C to invoke this and any failure is thrown as a tcl
exception. This means
that Tcl code can not build queues and that anything a tcl command does is
synchronous.

- please remove the translation from error code to string. OpenOCD is moving
*away* from using the return value as anything but a boolean success/failure.

It is not consistently reported/propagated or used, nor is it good API design
to propagate vital information in exceptions or error codes that the caller
needs to take action on. OpenOCD uses the return code to make some
decisions in a few places, but generally these are being weeded out
as time & occasion permits.

OpenOCD invokes LOG_ERROR() at the error site where more detailed
information about what went wrong is logged.

This should cut down the size of the patch considerably as well. Less code
is oftentimes a good sign.

- Having given it some thought, I would like to make it the C code's
responsibility
to invoke keep_alive and target_call_timer_callbacks_now(). Is there any
reason we can't do this? Really in a perfect world(or ultimately w/OpenOCD),
keep_alive & target_call_timer_callbacks_now() would go away. I've also
formulated a patch that makes keep_alive() invokoe
target_call_tiemr_callbacks_now().

- target->reset_halt should go away and be an argument to
target->assert_reset. At least from a tcl point of view. It gets a
little bit messier on the C end since deassert
also needs to know that the target is to be reset into the halted mode.



-- 
Øyvind Harboe
http://www.zylin.com/zy1000.html
ARM7 ARM9 XScale Cortex
JTAG debugger and flash programmer