On Thu, Mar 12, 2020 at 11:50:59AM +0100, Danil Kipnis wrote: > On Wed, Mar 11, 2020 at 8:03 PM Jason Gunthorpe <jgg@xxxxxxxx> wrote: > > > > On Wed, Mar 11, 2020 at 05:12:27PM +0100, Jack Wang wrote: > > > Add rtrs Makefile, Kconfig and also corresponding lines into upper > > > layer infiniband/ulp files. > > > > > > Signed-off-by: Danil Kipnis <danil.kipnis@xxxxxxxxxxxxxxx> > > > Signed-off-by: Jack Wang <jinpu.wang@xxxxxxxxxxxxxxx> > > > drivers/infiniband/Kconfig | 1 + > > > drivers/infiniband/ulp/Makefile | 1 + > > > drivers/infiniband/ulp/rtrs/Kconfig | 27 +++++++++++++++++++++++++++ > > > drivers/infiniband/ulp/rtrs/Makefile | 15 +++++++++++++++ > > > 4 files changed, 44 insertions(+) > > > create mode 100644 drivers/infiniband/ulp/rtrs/Kconfig > > > create mode 100644 drivers/infiniband/ulp/rtrs/Makefile > > > > How is this using ib_devices without having a struct ib_client ? > Hi Jason, > > After we resolved address using rdma_resolve_add() we access the > ib_device from rdma_cm_id. The ib_device has been registered on > rdma_cm module load in cma_init(). The handle device removal through > RDMA_CM_EVENT_DEVICE_REMOVAL. Hmm.. I actually don't know if this flow works OK in rdma cm or not.. Actually I expect if I look at rdma cm, I'll find it is broken.. However, it does sound like a reasonable approach, so nothing to do here. Jason
