On Sun, Jun 04, 2000 at 05:46:52PM +0100, Alan Cox wrote:
> > So is there no advantage of having something like that in the kernel, to
> > allow a program to queue up many buffers for the network code to handle
> > as it can, rather than either blocking on a send() or repeatedly making
> > kernel calls on a nonblocking socket until it manages to send all its
> > data?
> You can do that with a thread from user space. In fact a lot of systems do
> most of their AIO via threads in user space.
There are advantages to having this in the kernel which are detailed
on SGI's page (http://oss.sgi.com/projects/kaio/). I think Stephen
Tweedie is working on enhancements to Raw I/O which include async I/O
support in the kernel. They also provide significant performance
benefits.
--Chris
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Wed Jun 07 2000 - 21:00:19 EST