Texas Instruments TMS320C80, Руководство пользователя

Страница: 38 / 64

Technique #2: Using Pointers to Allocate RAM
2-8 TMS320C80 to TMS320C82 Software Compatibility User’s Guide
2.4 Technique #2: Using Pointers to Allocate RAM
Technique #1 can be augmented to take advantage of some of
the additional blocks of on-chip RAM that lie outside the popu-
lated regions common to the ’C80 and ’C82.
Technique #2 uses pointers to access on-chip RAM rather than
fixed address constants. With this technique, only the portion of
the software that actually allocates the RAM needs to recognize
the differences between the memory maps of the ’C80 and ’C82.
2.4.1 Advantages of Technique #2
This technique offers two main advantages:
-
Only the portion of the software that actually allocates the
RAM needs to recognize the differences between the
memory maps of the ’C80 and ’C82.
-
More available memory (compared to technique #1) can be
used on both devices.
2.4.2 Disadvantages of Technique #2
This technique has a two main disadvantages:
-
Technique #2 is more complicated than technique #1.
-
One RAM cannot be used since it is needed to store argu-
ments.
2.4.3 Implementing Technique #2
Implement this technique by writing software that performs three
tasks:
-
Access RAM through pointers rather than through fixed ad-
dress constants.
-
Use one block (or part of one block) of RAM as an argument
buffer.
-
Pass pointers to memory through the argument buffer and
use a command buffer to carry the location of the argument
buffer from the MP to each PP.
The software running on the MP should determine at system ini-
tialization time whether the device it is executing on is a ’C80 or
’C82. From this information, the MP should determine the loca-
tion of the additional RAM and then pass pointers to this RAM to