kern/96268: [socket] TCP socket performance drops by 3000% if packets are split at the first byte

[Tom Judge]

Jost Boekemeier wrote:
> Hi,
>
> from my point of view this issue can be closed. 
>
> TCP write/write/read sequences are bad on any operating system, it's just that other OS are a little bit smarter. -- I think Jon Nagle has had a proposal to fix/remove this unconditional delay, but I don't know if it has been implemented.
>
> Furthermore this problem has been fixed on application level. And I think Patrick van Staveren maintains a FreeBSD port which uses unix domain- instead of TCP socket communication.
>
>
> Regards,
> Jost Bökemeier
>   
Hi Jost,

I'm not sure if Patrick (I work with him) has a port using unix domain sockets in production.  I do know however that patches to PHP (That add the no delay socket option to the sockets API) where submitted after we found the work around for this issue and they where accepted.

And that we have a patched php java bridge client running in production.

My response was just to say that I have also seen this bug on a more recent release.

Regards

Tom