Re: [ECOS] deploying production systems using RedBoot as a bootloader

[Grant Edwards <grante at visi dot com>]

> Thanks for your response. Do you use the RedBoot 'fis'
> commands for doing field updates?

Not usually.  

The normal way to update was to wrap the new binary in a
program that knows about the FIS file system.  The wrapper
program (which includes the new binary as binary data) is
downloaded into RAM and started using one of the proprietary
admin protocols. (You could also load/run the wrapped program
with normal RedBoot commands.) The wrapper program then burns
the new binary into the FIS filesystem. Sort of like a
self-extracting zip file (that's not compressed), if you know
what I mean.

The standard method using a RedBoot "fis" commands and Y-modem
or TFTP is also supported, but it's a little too complex for
reliable field updates.

> You mentioned that you added 2 new protocols to use in place
> of the redboot cli. Are these protocols used to allow an
> administrator something other than a serial/telnet interface
> (like maybe a GUI running on a PC)?

Exactly.  The protocols themselves were defined and used for
previous products, so the were added primarily to maintain
backwards compatiblity.  The primary commands in the protocols
are used to:

 1) Query the device for identification (model, serial, etc.)
 2) Set the IP address
 3) Download a program into RAM and run it. 

We also provided a more machine-friendly version of the RedBoot
command line (accessed via serial or telnet) to be used to
automate things in production and test.  In that mode, RedBoot
commands aren't echoed, and every command replies with exactly
one line of CR/LF delimited text that starts with either a '-'
for failure or a '+' for success.

-- 
Grant Edwards
grante@visi.com

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss