[openib-general] [PATCH ] RFC IB/cm do not track remote QPN in timewait state
Michael S. Tsirkin
mst at mellanox.co.il
Tue Aug 29 10:50:14 PDT 2006
Quoting r. Sean Hefty <mshefty at ichips.intel.com>:
> Subject: Re: [openib-general] [PATCH ] RFC IB/cm do not track remote QPN in timewait state
>
> Michael S. Tsirkin wrote:
> >>I've thought about this too, and I think this may end up making the most sense.
> >>How would the driver determine how long the QP should remain in timewait,
> >
> >
> > Need to look into this - likely we can just add a call for that.
> > Roland?
>
> The Intel gen1 code passed this into the modify QP call for either reset or
> error. My question was more, does the driver have enough information to
> calculate the timewait duration based on other modify QP parameters? I don't
> know that you want the timewait duration determined by a userspace app.
>
> > There's no need for CM or any other entity to track timewait QP in this setup -
> > we can remove TIMEWAIT_EXIT event or pass it up immediately.
>
> The CM still needs to be able to respond to duplicate DREQs, so it will be
Correct of course - I should have said that CM will only need to track
he remote QP/ID and not the local QP, therefore there won't
be need for synchronisation beetween timewait in verbs layer and CM.
--
MST
More information about the general
mailing list