[openib-general] IPoIB fails attaching QP to mcast group

[Michael S. Tsirkin]

Quoting r. Or Gerlitz <ogerlitz at voltaire.com>:
> Subject: Re: [openib-general] IPoIB fails attaching QP to mcast group
> 
> Michael S. Tsirkin wrote:
> > Quoting r. Or Gerlitz <ogerlitz at voltaire.com>:
> 
> >> ib0: join completion for ff12:401b:ffff:0000:0000:0000:ffff:ffff (status 0)
> >> ib_mthca 0000:02:00.0: modify QP 3->3 returned status 10.
> >> ib0: failed to modify QP, ret = -22
> >> ib0: couldn't attach QP to multicast group ff12:401b:ffff:0000:0000:0000:ffff:ffff
> 
> > Looks like the QP is in error state, so modify QP fails.
> > Is this at all reproducible?
> > If so could you try with latest firmware please?
> 
> I am not following you, do you claim that the SW (IPoIB/MTHCA) consider 
> the QP to be in RTS but the FW/HW say that this QP is actually in error 
> state?

Seems like this.

> This happened today in endless loop on a system which i have played with 
> its IB link, specifically, i also saw the "recv port errors" counter was 
>   getting incremented. Once i have stopped/reloaded ipoib it does not 
> happen any more.
> 
> Or.
> 
> 

-- 
MST