On Tuesday 03 November 2009, Mariano Alvira wrote:
> arm7_9_write_core_regs did not update the core_cache and changes would
> get clobbered by arm7_9_full_context.

Hmm, that's what it's supposed to do.  The docs read:

arm7_9 write_core_reg num mode word                                    [Debug 
Command]
     This is intended for use while debugging OpenOCD; you probably shouldn’t 
use it.

     Writes a 32-bit word to register num (from 0 to 16) as used in the 
specified mode
     (where e.g. mode 16 is "user" and mode 19 is "supervisor"; the M4..M0 bits 
of the
     PSR). Registers 0..15 are the normal CPU registers such as r0(0), r1(1) 
... pc(15).
     Register 16 is the mode-specific SPSR, unless the specified mode is 
0xffffffff (32-bit
     all-ones) in which case register 16 is the CPSR. The write goes directly 
to the CPU,
     bypassing the register cache.

Making me wonder why the command still exists.  If no current
developers are using it, I'd be inclined to just remove it
in the 0.4.x cycle...

So that when you were typing

        arm7_9_write_core_regs 0 19 0xdeadbeef

you should instead have typed

        reg r0 0xdeadbeef

which would not have caused problems.  Does that work for you?

- Dave
_______________________________________________
Openocd-development mailing list
Openocd-development@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/openocd-development

Reply via email to