]> git.proxmox.com Git - mirror_ubuntu-kernels.git/commitdiff
Merge tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma
authorLinus Torvalds <torvalds@linux-foundation.org>
Fri, 5 Jun 2020 21:05:57 +0000 (14:05 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Fri, 5 Jun 2020 21:05:57 +0000 (14:05 -0700)
Pull rdma updates from Jason Gunthorpe:
 "A more active cycle than most of the recent past, with a few large,
  long discussed works this time.

  The RNBD block driver has been posted for nearly two years now, and
  flowing through RDMA due to it also introducing a new ULP.

  The removal of FMR has been a recurring discussion theme for a long
  time.

  And the usual smattering of features and bug fixes.

  Summary:

   - Various small driver bugs fixes in rxe, mlx5, hfi1, and efa

   - Continuing driver cleanups in bnxt_re, hns

   - Big cleanup of mlx5 QP creation flows

   - More consistent use of src port and flow label when LAG is used and
     a mlx5 implementation

   - Additional set of cleanups for IB CM

   - 'RNBD' network block driver and target. This is a network block
     RDMA device specific to ionos's cloud environment. It brings strong
     multipath and resiliency capabilities.

   - Accelerated IPoIB for HFI1

   - QP/WQ/SRQ ioctl migration for uverbs, and support for multiple
     async fds

   - Support for exchanging the new IBTA defiend ECE data during RDMA CM
     exchanges

   - Removal of the very old and insecure FMR interface from all ULPs
     and drivers. FRWR should be preferred for at least a decade now"

* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma: (247 commits)
  RDMA/cm: Spurious WARNING triggered in cm_destroy_id()
  RDMA/mlx5: Return ECE DC support
  RDMA/mlx5: Don't rely on FW to set zeros in ECE response
  RDMA/mlx5: Return an error if copy_to_user fails
  IB/hfi1: Use free_netdev() in hfi1_netdev_free()
  RDMA/hns: Uninitialized variable in modify_qp_init_to_rtr()
  RDMA/core: Move and rename trace_cm_id_create()
  IB/hfi1: Fix hfi1_netdev_rx_init() error handling
  RDMA: Remove 'max_map_per_fmr'
  RDMA: Remove 'max_fmr'
  RDMA/core: Remove FMR device ops
  RDMA/rdmavt: Remove FMR memory registration
  RDMA/mthca: Remove FMR support for memory registration
  RDMA/mlx4: Remove FMR support for memory registration
  RDMA/i40iw: Remove FMR leftovers
  RDMA/bnxt_re: Remove FMR leftovers
  RDMA/mlx5: Remove FMR leftovers
  RDMA/core: Remove FMR pool API
  RDMA/rds: Remove FMR support for memory registration
  RDMA/srp: Remove support for FMR memory registration
  ...

1  2 
MAINTAINERS
drivers/infiniband/core/rdma_core.c
drivers/infiniband/ulp/ipoib/ipoib_main.c
drivers/net/ethernet/qlogic/qed/qed_rdma.c
drivers/net/ethernet/qlogic/qed/qed_rdma.h
drivers/nvme/target/rdma.c
fs/sysfs/file.c
include/linux/qed/qed_rdma_if.h
include/rdma/uverbs_std_types.h
net/smc/smc_ib.c

diff --cc MAINTAINERS
Simple merge
Simple merge
Simple merge
diff --cc fs/sysfs/file.c
Simple merge
Simple merge
Simple merge
index f0a5064bf9bd756d21c0bfaecdfafd63a4fa02b2,2fad5f3fe09392aa67b6c5f92351d721cc7f7d57..562a52d01ad161749cb9bbd5a9398416bd292e55
@@@ -582,18 -579,11 +582,19 @@@ static int smc_ib_add_dev(struct ib_dev
             i++) {
                set_bit(i, &smcibdev->port_event_mask);
                /* determine pnetids of the port */
 -              smc_pnetid_by_dev_port(ibdev->dev.parent, i,
 -                                     smcibdev->pnetid[i]);
 +              if (smc_pnetid_by_dev_port(ibdev->dev.parent, i,
 +                                         smcibdev->pnetid[i]))
 +                      smc_pnetid_by_table_ib(smcibdev, i + 1);
 +              pr_warn_ratelimited("smc:    ib device %s port %d has pnetid "
 +                                  "%.16s%s\n",
 +                                  smcibdev->ibdev->name, i + 1,
 +                                  smcibdev->pnetid[i],
 +                                  smcibdev->pnetid_by_user[i] ?
 +                                   " (user defined)" :
 +                                   "");
        }
        schedule_work(&smcibdev->port_event_work);
+       return 0;
  }
  
  /* callback function for ib_unregister_client() */