[Click] [RFC] [PATCH] PollDevice/ToDevice improvement

Adam Greenhalgh a.greenhalgh at cs.ucl.ac.uk
Fri Nov 16 15:07:49 EST 2007


Hi folks,

Has any one got performance results for 60byte packets using Joowoo's
new patch against a the polling driver ?

Adam

On 16/11/2007, Joonwoo Park <joonwpark81 at gmail.com> wrote:
> 2007/11/16, springbo at cs.wisc.edu <springbo at cs.wisc.edu>:
> > Hello Joonwoo,
> >
> > The new patch compiled without any warnings on x86_64.
> >
> > After an hour of use with an Intel PRO/1000 PT Dual Port Server Adapter
> > NIC I haven't seen any major problems with the patch so far. It seems to
> > be working fine.
>
> That's good!
>
> >
> > I did see a Tx Unit Hang when shutting down kclick:
> > e1000: eth2: e1000_clean_tx_irq: Detected Tx Unit Hang
> >  Tx Queue             <0>
> >  TDH                  <cf>
> >  TDT                  <cf>
> >  next_to_use          <cf>
> >  next_to_clean        <62>
> > buffer_info[next_to_clean]
> >  time_stamp           <100258906>
> >  next_to_watch        <62>
> >  jiffies              <100284bf5>
> >  next_to_watch.status <1>
> >
>
> Thanks for it. I looks like we need to clean tx ring to stop click.
> I'll check this.
>
> > I haven't crunched the numbers for performance yet, but I did notice the
> > machine seemed to be much more responsive.
> >
>
> It sounds hopeful. :-)
> Thanks so much.
>
> joonwpark.tistory.com
> Joonwoo
> _______________________________________________
> click mailing list
> click at amsterdam.lcs.mit.edu
> https://amsterdam.lcs.mit.edu/mailman/listinfo/click
>


More information about the click mailing list