Re: PATCH: Raw device IO for 2.1.131

David S. Miller (davem@dm.cobaltmicro.com)
Mon, 14 Dec 1998 03:09:03 -0800


From: Jes Sorensen <Jes.Sorensen@cern.ch>
Date: 14 Dec 1998 11:10:14 +0100

SGI managed to get zero copy right for write() on a socket and it
works great, I'd love having Linux do the same.

Here is where I disagree with you, sendfile() is the right answer to
pushing data to the wire, SGI's solution is not right and adds
complexity to the kernel where it doesn't need to exist.

For receive and writes to a disk, sendfile() is not the answer.

So you have a foo_linux.c module which uses sendfile for socket
writes in order to take advantage of sendfile, if this is too much to
ask then sorry. I never buy the argument from people saying that "if
you add a new API that I might want to use for performance reasons, I
may have to abstract out some of my application code and make it more
portable as a side effect!" oh horrors....

Later,
David S. Miller
davem@dm.cobaltmicro.com

-
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/