On Mon, Jul 09, 2018 at 10:02:32AM +0300, Leon Romanovsky wrote: > On Sun, Jul 08, 2018 at 03:27:03PM +0300, Yuval Bason wrote: > > From: Yuval Bason <Yuval.Bason@xxxxxxxxxx> > > > > Enable SRQ's for rdma-core. > > > > This patch will be applied in rdma-core, so there is no need to mention > "rdma-core" in commit message. > > > SRQ rdma-core implementation is pretty straight forward, except perhaps the > > Way that the driver udpates FW producers: they are updated using a mapped buffer > > that FW reads, and not doorbells Like RQ / SQ. > > > > Signed-off-by: Yuval Bason <yuval.bason@xxxxxxxxxx> > > Signed-off-by: Michal Kalderon <michal.kalderon@xxxxxxxxxx> > > --- > > kernel-headers/rdma/qedr-abi.h | 15 ++ > > providers/qedr/qelr.h | 43 ++++++ > > providers/qedr/qelr_abi.h | 2 + > > providers/qedr/qelr_main.c | 7 + > > providers/qedr/qelr_verbs.c | 337 +++++++++++++++++++++++++++++++++++++++++ > > providers/qedr/qelr_verbs.h | 9 ++ > > providers/qedr/rdma_common.h | 1 + > > 7 files changed, 414 insertions(+) > > <...> > > @@ -2018,6 +2352,9 @@ void qelr_async_event(struct ibv_async_event *event) > > case IBV_EVENT_COMM_EST: > > case IBV_EVENT_QP_LAST_WQE_REACHED: > > break; > > + case IBV_EVENT_SRQ_LIMIT_REACHED: > > + case IBV_EVENT_SRQ_ERR: > > If you are here, please put /* fail through */ comment for gcc-7. > This comment is wrong, there is no need in "/* fail through */" if no code is executed under case. Thanks
Attachment:
signature.asc
Description: PGP signature
