Hi,

I had a very fast look at the latest beta2 and I'd like to address some
points, apoligizing if I got something wrong:

- Are tasks bounded to a single cpu? If so they cannot run on a cluster
or on all. Is it not too heavy a constraint?

- following that there is no more, esplicitely, the concept of a
rt_current task. For me it is not a good idea. Linux kernel has kept the
concept  allowing the reuse of much of the UP code, it is a lot better
choice.

- SMP atomicity is insured by a plain spinlock. So if people nest the
atomicity call they'll find them working on UP but get stucked on SMP?
If this is true why not inheredit the global cli/sti.... idea from
Linux? Even if tasks are bound on a single cpu I expect user can have
needs to nest atomic operation or have an easy way to synchronize across
cpus, without setting specific spinlocks.

I'd like some comments on this design issues from the develeppers, if I
can ask for them.

Ciao, Paolo.
--- [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/

Reply via email to