Block Headers - Analog Devices VisualDSP++ 3.5 Manual

This first header is followed by the regular boot stream, that is, a series of
headers and data blocks. Most headers are followed by corresponding
blocks of data, but some headers indicate regions of memory that need to
be "zero-filled" and are not followed by data blocks.

Block Headers

Each block header consists of four or six 16-bit words:
• The first word consists of a flag that indicates whether the follow-
ing block of data is a 24-bit or 16-bit payload or zero-initialized
data. The flag uniquely identifies the last block that needs to be
transferred.
• The second word contains the lower 16 bits of the 24-bit start
address for data loading (destination). The first octet is the 8 LSBs,
followed by the next most significant bits (
• The third word contains the upper-most 8 bits of the 24-bit desti-
nation address, padded (suffixed) with one byte of zeros.
• The fourth word contains the payload's word count. Similar to the
address, the first octet is the 8 LSBs, and the second octet is the 8
MSBs.
An extra word appears when a
accuracy. This fifth word (also a 16-bit field) is the CRC-16
the header, and the data block immediately follows it. The CRC
word is optional. Activate the
the
-checksum
The header is buffered with a dummy word of zeros to ease the EMI
addressing issue.
EPROM booting can be performed in an 8-bit or 16-bit scenario. This
information is controlled by the
ded in the boot stream. Unlike non-boot mode, bus width is not
VisualDSP++ Loader Manual
for 16-Bit Processors
checksum
checksum
command-line switch
"-width #"
ADSP-219x DSP Loader/Splitter
), and so on.
15–8
function is used to verify booting
by running the loader utility with
(on page 3-21).
switch and is finally embed-
for
checksum
checksum
3-5