Hi,u

I just merged a patch into SVN which removes some probably never used
feature of RTnet: socket creation and destruction in hard RT context.

If you never played with the module parameter rtskb_cache_size, you
don't need to worry at all. If you did, you either wrote a kernel-hosted
RT application and called [rtdev_]socket() from real-time tasks, or you
noticed that creating sockets from userspace RT tasks sometimes failed -
precisely always then when the caller happened to run in RT (Xenomai:
"primary") context.

In the former case (kernel application), you should now move all your
socket setup/cleanup code into module_init/cleanup context. In the
latter case, the real-time Linux extension now simply switch the caller
of [rtdev_]socket() into the right context: non-RT. Thus you don't need
to worry about the rtskb_cache anymore, all buffers for your socket are
allocated from Linux pools. The same goes for extending/shrinking those
socket buffer pools: all happens in non-RT now.

All in all, this removal is expected to simplify the use of RTnet,
specifically in userspace applications.

Jan

Attachment: signature.asc
Description: OpenPGP digital signature

-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
_______________________________________________
RTnet-users mailing list
RTnet-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/rtnet-users

Reply via email to