On Sun, Sep 27, 2009 at 7:28 PM, David Brownell <davi...@pacbell.net> wrote: > On Sunday 27 September 2009, Ųyvind Harboe wrote: >> I can't find the place in the ft2232 code where the state is set >> to TAP_RESET when srst pulls trst under a reset... > > I'd have expected that to be in common jtag/core.c code ... > > >> How does the attached patch look? >> >> rlink.c, bitbang.c, bitq.c and zy1000 all have this code >> path right. > > .... instead of needing to modify every driver. In this > case there's no obstacle to fixing that once in the core. > (Instead of N times in each driver...)
I don't think you can or should do this in the core driver. core.c has no way of knowing and shouldn't know how each driver tracks the current state. The current TAP state can be tracked by hardware and it would add software overhead(yes, this *does* matter for embedded implementations) to track the TAP state. The current solution is relatively straightforward, even if it is a bit of copy & paste. -- Øyvind Harboe http://www.zylin.com/zy1000.html ARM7 ARM9 ARM11 XScale Cortex JTAG debugger and flash programmer _______________________________________________ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
