[openfabrics-ewg] OFED procedures

Scott Weitzenkamp (sweitzen) sweitzen at cisco.com
Sun Oct 15 16:54:49 PDT 2006


I went to do this, and I see I have to get this change into my copy of
the OFED openib src RPM from rc7.

How do I do that?

Scott Weitzenkamp
SQA and Release Manager
Server Virtualization Business Unit
Cisco Systems
 

> -----Original Message-----
> From: Michael S. Tsirkin [mailto:mst at mellanox.co.il] 
> Sent: Monday, October 09, 2006 11:06 PM
> To: Scott Weitzenkamp (sweitzen)
> Cc: Aviram Gutman; Open Fabrics; Eitan Zahavi
> Subject: Re: [openfabrics-ewg] OFED procedures
> 
> In that the SM does not obey the MTU selector? It could.
> 
> Can you try going into cm.c cm_init_qp_rtr_attr, and
> printk the value of qp_attr->path_mtu near after it is set?
> 
> 
> Quoting r. Scott Weitzenkamp (sweitzen) <sweitzen at cisco.com>:
> > Subject: RE: [openfabrics-ewg] OFED procedures
> > 
> > tavor_quirk=1 with rdma_cm didn't help, could this be a 
> Cisco SM issue?
> > 
> > Scott Weitzenkamp
> > SQA and Release Manager
> > Server Virtualization Business Unit
> > Cisco Systems
> >  
> > 
> > > -----Original Message-----
> > > From: Michael S. Tsirkin [mailto:mst at mellanox.co.il] 
> > > Sent: Wednesday, October 04, 2006 12:57 PM
> > > To: Scott Weitzenkamp (sweitzen)
> > > Cc: Aviram Gutman; Open Fabrics; Eitan Zahavi
> > > Subject: Re: [openfabrics-ewg] OFED procedures
> > > 
> > > Quoting Scott Weitzenkamp (sweitzen) <sweitzen at cisco.com>:
> > > > Subject: RE: [openfabrics-ewg] OFED procedures
> > > > 
> > > > PCI-X max throughput is still about 30% below what is was 
> > > before.  More
> > > > details are in http://openib.org/bugzilla/show_bug.cgi?id=108.
> > > 
> > > Oh, that will be the tavor MTU quirk.
> > > You have 2 options:
> > > - activate the tavor quirk workaround in opensm (if you 
> use opensm)
> > >   Eitan, can you pls provide guidance on how to enable?
> > > - activate the tavor quirk workaround in cma
> > >   set tavor_quirk=1 module parameter for the rdma_cm module
> > > 
> > > -- 
> > > MST
> > > 
> > 
> 
> -- 
> MST
> 




More information about the ewg mailing list