[openib-general] Re: Re: CMA deadlock

Michael S. Tsirkin mst at mellanox.co.il
Mon Apr 3 13:55:13 PDT 2006


Quoting r. Sean Hefty <mshefty at ichips.intel.com>:
> Subject: Re: Re: CMA deadlock
> 
> Michael S. Tsirkin wrote:
> >The current approach seems racy: what if e.g. listen call is in progress?  
> >It
> >looks at id_priv->cm_id.ib which seemingly might change under its feet.
> 
> There shouldn't be any races.  Callbacks from the IB CM are serialized, so 
> we're in the only callback that can occur.  And if the user destroys the 
> rdma_id by returning a non-zero value in the callback, then they cannot be 
> using that same id in any other call.

Sounds sane. OK, please commit the patch you've sent and I'll test tomorrow.

-- 
Michael S. Tsirkin
Staff Engineer, Mellanox Technologies



More information about the general mailing list