Hello all,
as practice shows, RT-fifos can reasonably be used only in one
direction. I conclude that from a simple test, where I tried
to use one fifo for transmission from user space to RTL and
back (my handler tried to immediately return the received
message). After the first message from user space to RTL the
RTL application did loop forever, since it always got it's own
message. Is that right?
If so, could it easily be changed? I would like to write a
device driver using RTL, which can be accessed from a user
space application. And like usual behaviour, the user space
application would like to write and read on the same device.
Needing two fifos would be less comforatble, furthermore with
this bidirectional capability every standard driver could be
replaced by simply creating the RTL workalike and giving the
correct name to the fifo.
Only other possibility for drivers with realtime requirements
would be to have a standard kernel driver and have that commu-
nicate with a RTL driver. Would complicate things and has the
problem of communication between Linux kernel and RTL.
Thanks for your help,
Markus
---
Sent through Global Message Exchange - http://www.gmx.net
--- [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/
