On Tuesday 07 July 2009, Øyvind Harboe wrote:
> > I can believe the chip requires "reset_config srst_pulls_trst".
> >
> > But I have a hard time believing "srst_only" is anything but
> > a board-specific constraint.
> 
> The TRST is tied to SRST on the *inside* of the chip if I understood
> correctly.

Fine, but that's beside the point.


> I'm not sure if there is a difference between "srst_pulls_trst srst_only"
> and "srst_pulls_trst" from the arm926ejs codepaths point of view.

That seems kind of buglike.  But, I'm moderately convinced
some of the reset problems are going to require lowlevel
surgery to resolve them.  The code is just not factored well
enough to support the variety of real-world reset models we
now observe.  It originally handled just a few models...

If there's no SRST, then any logic depending on it should
obviously never kick in.  And other commands should be used
to reset a given target.  (E.g. through a DAP or JRC.)

But it should also be possible to say that *IF* a board gets
reset using SRST, that has specific consequences for JTAG
on a given chip.

Note that's not quite the same as saying that the board
has screwey wiring that connects TRST and SRST.  What you
summarize above applies to a single chip.  But there can
also be board-level behaviors ... consider a board with
an i.MX27 plus another chip that doesn't couple those
behaviors internally.  Issue SRST ... and one chip will
have undisturbed JTAG/debug state, while the i.MX27 is
going to need some re-init.

- Dave

_______________________________________________
Openocd-development mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/openocd-development

Reply via email to