Re: [PATCH] tcp: splice as many packets as possible at once

[Willy Tarreau <w@1wt.eu>]

On Sat, Jan 10, 2009 at 02:01:24AM +0300, Evgeniy Polyakov wrote:
> On Fri, Jan 09, 2009 at 11:50:10PM +0100, Willy Tarreau (w@1wt.eu) wrote:
> > Well, I'm now trying to educate myself on the send part. It's still
> > not very clear to me and I'd like to understand a little bit better
> > why we have this corruption problem and why there is a difference
> > between sending segments from memory and sending them from another
> > socket where they were already waiting.
> 
> printks are the best choice, since you will get exactly what you are
> looking for instead of deciphering what developer or code told you.

I know, but as I told in an earlier mail, it was not even clear to
me what function were called. I have made guesses but that's quite
hard when you don't know the entry point. I think I'm not too far
from having discovered the call chain. Understanding it will be
another story, of course ;-)

> > I think I'll put printks everywhere and see what I can observe.
> > Knowing about the GSO/SG workaround already helps me enable/disable
> > the bug.
> 
> I wish I could also be capable to disable the bugs... :)

Me too :-)
Here we have an opportunity to turn it on/off, let's take it !

Willy