Hi Linus, We have 5 small fixes for this pull request. One is a performance regression, so not necessarily strictly a fix, but it was small and reasonable and claimed to avoid thrashing in the scheduler, so I took it. The remaining are all legitimate fixes that match the "we take fixes any time" criteria. Here's the boilerplate: The following changes since commit 7bca603a69c0c239654a8f0bcb99e1a60b30040c: RDMA/mlx5: Initialize return variable in case pagefault was skipped (2018-11-29 15:16:45 -0700) are available in the Git repository at: git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma.git tags/for-linus for you to fetch changes up to 37fbd834b4e492dc41743830cbe435f35120abd8: IB/core: Fix oops in netdev_next_upper_dev_rcu() (2018-12-12 12:14:49 -0500) ---------------------------------------------------------------- Pull request for 4.20-rc - One performance regression for hfi1 - One kasan fix for hfi1 - A couple mlx5 fixes - A core oops fix ---------------------------------------------------------------- Artemy Kovalyov (1): IB/mlx5: Fix implicit ODP interrupted page fault Mark Zhang (1): IB/core: Fix oops in netdev_next_upper_dev_rcu() Michael J. Ruhl (1): IB/hfi1: Fix a latency issue for small messages Piotr Stankiewicz (1): IB/hfi1: Fix an out-of-bounds access in get_hw_stats Yishai Hadas (1): IB/mlx5: Block DEVX umem from the non applicable cases drivers/infiniband/core/roce_gid_mgmt.c | 3 +++ drivers/infiniband/hw/hfi1/chip.c | 3 ++- drivers/infiniband/hw/hfi1/hfi.h | 2 ++ drivers/infiniband/hw/hfi1/qp.c | 7 +++++++ drivers/infiniband/hw/hfi1/verbs.c | 2 +- drivers/infiniband/hw/mlx5/devx.c | 4 +++- drivers/infiniband/hw/mlx5/odp.c | 9 ++++----- 7 files changed, 22 insertions(+), 8 deletions(-) -- Doug Ledford GPG KeyID: B826A3330E572FDD Key fingerprint = AE6B 1BDA 122B 23B4 265B 1274 B826 A333 0E57 2FDD