Re: zero-copy TCP

From: Chris Wedgwood (cw@f00f.org)
Date: Tue Sep 05 2000 - 14:45:05 EST


On Sun, Sep 03, 2000 at 08:46:15PM +0100, Alan Cox wrote:

    You are perfectly at liberty to send invalidly checksummed
    packets. Its a design decision. The remote simply discards the
    frame

Not only that -- from outside the box it's often impossible to tell
why a certain packet has a broken checksum; it might be software,
hardware a stray event or interference for the Redmond Mind Control
Rays.

Right now; bad checksums occur on most networks[1] fairly often
anyhow; I guess Ingo was trying to say that we should make efforts
not to increase this number if at all possible?

  --cw

[1] Well, reasonable sized networks. I guess it goes O(n) with the
    number of hosts in the absence of something really broken.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Thu Sep 07 2000 - 21:00:23 EST