From: Aharon Landau Date: Tue, 15 Feb 2022 17:55:30 +0000 (+0200) Subject: RDMA/mlx5: Fix the flow of a miss in the allocation of a cache ODP MR X-Git-Tag: v5.16.19~473 X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=8bb42130f90cbaf777da534e73e727b913bff908;p=thirdparty%2Fkernel%2Fstable.git RDMA/mlx5: Fix the flow of a miss in the allocation of a cache ODP MR [ Upstream commit 2f0e60d5e9f96341a0c8a01be8878cdb3b29ff20 ] When an ODP MR cache entry is empty and trying to allocate it, increment the ent->miss counter and call to queue_adjust_cache_locked() to verify the entry is balanced. Fixes: aad719dcf379 ("RDMA/mlx5: Allow MRs to be created in the cache synchronously") Link: https://lore.kernel.org/r/09503e295276dcacc92cb1d8aef1ad0961c99dc1.1644947594.git.leonro@nvidia.com Signed-off-by: Aharon Landau Signed-off-by: Leon Romanovsky Signed-off-by: Jason Gunthorpe Signed-off-by: Sasha Levin --- diff --git a/drivers/infiniband/hw/mlx5/mr.c b/drivers/infiniband/hw/mlx5/mr.c index 157d862fb8642..2910d78333130 100644 --- a/drivers/infiniband/hw/mlx5/mr.c +++ b/drivers/infiniband/hw/mlx5/mr.c @@ -585,6 +585,8 @@ struct mlx5_ib_mr *mlx5_mr_cache_alloc(struct mlx5_ib_dev *dev, ent = &cache->ent[entry]; spin_lock_irq(&ent->lock); if (list_empty(&ent->head)) { + queue_adjust_cache_locked(ent); + ent->miss++; spin_unlock_irq(&ent->lock); mr = create_cache_mr(ent); if (IS_ERR(mr))