On Mon, Sep 14, 2009 at 08:35:32PM +0200, Øyvind Harboe wrote: > On Mon, Sep 14, 2009 at 7:42 PM, David Brownell <davi...@pacbell.net> wrote: > > On Monday 14 September 2009, Ųyvind Harboe wrote: > >> I suspect there are JTAG controllers that do not communicate when > >> SRST is asserted, but do not pull TRST. > >> > >> srst_pulls_trst will kinda work on these devices i.MX27(??). > >> > >> The first thought that strikes me here is to split out srst_pulls_trst into > >> two flags: pulls trst and no JTAG communication during srst.... > > > > What code do we have which wants JTAG chitchat during SRST? > > I thought arm7/9 did this, but I was wrong. I'm wondering if > XScale needs this. > > Turns out after some more testing that i.MX27 does appear to > pull trst during srst.
I thought the desired reset procedure at least on ARM9 is: - pull SRST and TRST (TRST not only resets the TAP but also EmbeddedICE) - release TRST - program EmbeddedICE to break on reset vector - release SRST There are SoCs where SRST gates the ARM core clock, thus you cannot program EmbeddedICE while SRST is asserted; then you need: - pull SRST and TRST - release SRST and TRST - program EmbeddedICE to break on reset vector - pull SRST again - release SRST Johannes _______________________________________________ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development