Sean Hefty wrote: > I think there's a related issue in the ib_cm, which also sets > IB_ACCESS_LOCAL_WRITE. The above code is executed when the user calls > rdma_create_qp(). The ib_cm routine is executed when connecting the QP, > which > will overwrite these settings. I think we'll want to change both places to > get > the desired result.
I've pulled in your patch, along with a similar patch to the ib_cm. I'll run through some tests tomorrow to verify that nothing broke. Unless this is causing a serious issue, though, I will request this fix for 2.6.20. - Sean _______________________________________________ openib-general mailing list [email protected] http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
