[EMAIL PROTECTED] wrote:
> Other question is, If I can access files from a RT Module?.
Julian,
As another thread in this message list has pointed out: "The short
answer is .. not yet."
You'll need to break your application up into initialization (INIT),
dynamic (DYN) and post-processsing (POST) phases. Direct writes to the
hard disk are not allowed during the DYN (i.e., realtime) phase. INIT
and POST portions can usually be handled by regular Linux processes that
send data to/from your RTLinux DYN tasks via FIFO's or shared memory.
Many applications do not need to write data to a file during DYN
operations--they just store results in memory and hand them off to the
regular Linux POST routine when the program has finished. If you need
to stream results to disk, you can create a regular Linux 'data catcher'
application that reads data from a realtime FIFO (or shared memory) and
writes it to disk.
If your realtime tasks require data from files, you can write a Linux
"data server" to send that data to your RT tasks. However, you should
be aware that the delivery time of this data cannot be guaranteed under
RTLinux. It is not possible to say: "Get me the data from file
my_dat.dat every 10 ms." It might get there within 10 ms, but it might
not. If on-time, _every_ time delivery of data from files is a
requirement for you, you might want to consider other RTOSes.
If you can't rework your application to avoid file access during the
realtime phase, and if you don't need to meet really tight deadlines,
you might consider just programming to the regular POSIX realtime
interface. Your guaranteed latency will be greater, but you'll have the
convenience of being able to do regular file I/O..
Hope that helps,
- Dave
--- [rtl] ---
To unsubscribe:
echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR
echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED]
----
For more information on Real-Time Linux see:
http://www.rtlinux.org/~rtlinux/
