mirror of
https://github.com/hsoft/collapseos.git
synced 2024-09-07 12:28:48 +10:00
also, verify all 3 first bytes of SPI commands. I'm not sure why I wasn't doing that, probably because I was getting a lot of AVR err and thought that only 2 bytes of the cmd were echoed. But now, with a reliable SPI setup, verifying 3 bytes seems to work.
113 lines
4.3 KiB
Plaintext
113 lines
4.3 KiB
Plaintext
# Programming AVR chips
|
|
|
|
(In this documentation, you are expected to have an AVR binary
|
|
ready to send. To assemble an AVR binary from source, see
|
|
asm.txt)
|
|
|
|
To program AVR chips, you need a device that provides the SPI
|
|
protocol. The device built in the rc2014/sdcard recipe fits the
|
|
bill. Make sure you can override the SPI clock because the sys-
|
|
tem clock will be too fast for most AVR chips, which are usually
|
|
running at 1MHz. Because the SPI clock needs to be a 4th of
|
|
that, a safe frequency for SPI communication would be 250kHz.
|
|
|
|
# The programmer device
|
|
|
|
The AVR programmer device is really simple: Wire SPI connections
|
|
to proper AVR pins as described in the MCU's datasheet. Note
|
|
that this device will be the same as the one you'll use for any
|
|
modern SPI-based AVR programmer, with RESET replacing SS.
|
|
|
|
This device should have an on/off switch that controls the
|
|
chip's power for a very simple reason: Because we can't control
|
|
what's on the chip, it could mess up your whole SPI bus when
|
|
RESET is not held low. This means that as long as it's connected
|
|
and powered, it is likely to mess up your other devices, such as
|
|
the SD card.
|
|
|
|
You could put the AVR chip behind a buffer to avoid this, but
|
|
an on/off switch also does the trick and satisfies the low-tech
|
|
lover in you.
|
|
|
|
# Programming software
|
|
|
|
The AVR programming code is at B160.
|
|
|
|
Before you begin programming the chip, the device must be desel-
|
|
ected. Ensure with "0 (spie)".
|
|
|
|
Then, you initiate programming mode with "asp$", and then issue
|
|
your commands.
|
|
|
|
Each command will verify that it's in sync, that is, that its
|
|
3rd exchange echoes the byte that was sent in the 2nd exchange.
|
|
If it doesn't, the command aborts with "AVR err".
|
|
|
|
# Ensuring reliability
|
|
|
|
The reliability of your communication depends a lot on the
|
|
soundness of your SPI relay design. If it's good, you will sel-
|
|
dom see those "AVR err".
|
|
|
|
However, there are worse things than "AVR err": wrong data. Sync
|
|
checks ensure communication reliability at every command, but
|
|
in the case of commands getting data, you might be out-of-sync
|
|
when you receive your result without knowing it! To ensure that
|
|
you're still in sync, you need to issue a command, which might
|
|
spit "AVR err". If it does, your previous result is unreliable.
|
|
|
|
Here's an example word that reliably prints the high fuse value
|
|
from SPI devid 1:
|
|
|
|
: get 1 asp$ asprdy aspfh@ asprdy .x 0 (spie) ;
|
|
|
|
Another very important matter is clock speed. As mentioned
|
|
above, the safe clock speed is 250kHz. If you use the SPI design
|
|
in rc2014/sdcard recipe, this means that your input clock speed
|
|
can theoretically be 500kHz because the '161 divides it by 2.
|
|
|
|
In practice, however, you can't really do that because depending
|
|
on the timing of your SPI write, the first "bump" of the SPI
|
|
clock might end up being nearly 500kHz, which will result in oc-
|
|
casional communication errors.
|
|
|
|
The simplest and safest way to avoid this is to reduce your
|
|
raw input clock by 2, which will reduce your effective communi-
|
|
cation speed by 2. There certainly are options allowing you to
|
|
keep optimal speed, but they're significantly more complex than
|
|
accepting slower speed.
|
|
|
|
# Access fuses
|
|
|
|
You get/set they values with "aspfx@/aspfx!", x being one of "l"
|
|
(low fuse), "h" (high fuse), "e" (extended fuse).
|
|
|
|
# Access flash
|
|
|
|
Writing to AVR's flash is done in batch mode, page by page. To
|
|
this end, the chip has a buffer which is writable byte-by-byte.
|
|
|
|
Writing to the flash begins with a call to asperase, which
|
|
erases the whole chip. It seems possible to erase flash page-by-
|
|
page through parallel programming, but the SPI protocol doesn't
|
|
expose it, we have to erase the whole chip. Then, you write to
|
|
the buffer using aspfb! and then write to a page using aspfp!.
|
|
Example to write 0x1234 to the first byte of the first page:
|
|
|
|
asperase 0x1234 0 aspfb! 0 aspfp!
|
|
|
|
Please note that aspfb! deals with *words*, not bytes. If, for
|
|
example, you want to hook it to A!*, make sure you use AMOVEW
|
|
instead of AMOVE. You will need to create a wrapper word around
|
|
aspfb! that divides dst addr by 2 because AMOVEW use byte-based
|
|
addresses but aspfb! uses word-based ones. You also have to make
|
|
sure that A@* points to @ (or another word-based fetcher)
|
|
instead of its default value of C@.
|
|
|
|
# Access EEPROM
|
|
|
|
Accessing EEPROM is simple and is done byte-by-byte with words
|
|
aspe@ and aspe!. Example:
|
|
|
|
0x42 0 aspe! 0 aspe@ .x ( prints 42 )
|