David> From: Jes Sorensen <Jes.Sorensen@cern.ch> Date: 14 Dec 1998
David> 14:10:08 +0100
David> Here is where I disagree with you, sendfile() is the right
David> answer to pushing data to the wire, SGI's solution is not right
David> and adds complexity to the kernel where it doesn't need to
David> exist.
> Provided that data is read directly from disk and put on the
> wire, yes.
David> And you were originally talking about fileserving and how Linux
David> lacks there, and this is why I made the statement I did. Now
David> you run off and say it is yet another application of data I/O
David> that we need something other than sendfile for.
Ok I should have mentioned that example in my previous posting. I
agree that sendfile() solves some of the problems, but without having
some sort of standard for it (afaik it's not accepted by POSIX or
anything, I could be wrong, haven't checked properly) I will have a
hard time telling people to use it - I will try though.
Anyway sendfile() is a special case and doesn't solve the whole
problem and as such I still think zero copy on write() is justified.
Sorry for the confusion.
Jes
-
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/