| Message ID | 88699500f690dff1c1852c1ddb71f8a1cc8b956e.1733233480.git.leonro@nvidia.com (mailing list archive) |
|---|---|
| State | Awaiting Upstream |
| Headers | show |
| Series | [mlx5-next] RDMA/mlx5: Enforce same type port association for multiport RoCE | expand |
| Context | Check | Description |
|---|---|---|
| netdev/tree_selection | success | Not a local patch |
On 12/3/2024 2:45 PM, Leon Romanovsky wrote: > From: Patrisious Haddad <phaddad@nvidia.com> > > Different core device types such as PFs and VFs shouldn't be affiliated > together since they have different capabilities, fix that by enforcing > type check before doing the affiliation. > > Fixes: 32f69e4be269 ("{net, IB}/mlx5: Manage port association for multiport RoCE") > Reviewed-by: Mark Bloch <mbloch@nvidia.com> > Signed-off-by: Patrisious Haddad <phaddad@nvidia.com> > Signed-off-by: Leon Romanovsky <leonro@nvidia.com> > --- > drivers/infiniband/hw/mlx5/main.c | 6 ++++-- > include/linux/mlx5/driver.h | 6 ++++++ > 2 files changed, 10 insertions(+), 2 deletions(-) > > diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c > index bc7930d0c564..c2314797afc9 100644 > --- a/drivers/infiniband/hw/mlx5/main.c > +++ b/drivers/infiniband/hw/mlx5/main.c > @@ -3639,7 +3639,8 @@ static int mlx5_ib_init_multiport_master(struct mlx5_ib_dev *dev) > list_for_each_entry(mpi, &mlx5_ib_unaffiliated_port_list, > list) { > if (dev->sys_image_guid == mpi->sys_image_guid && > - (mlx5_core_native_port_num(mpi->mdev) - 1) == i) { > + (mlx5_core_native_port_num(mpi->mdev) - 1) == i && > + mlx5_core_same_coredev_type(dev->mdev, mpi->mdev)) { > bound = mlx5_ib_bind_slave_port(dev, mpi); > } > > @@ -4785,7 +4786,8 @@ static int mlx5r_mp_probe(struct auxiliary_device *adev, > > mutex_lock(&mlx5_ib_multiport_mutex); > list_for_each_entry(dev, &mlx5_ib_dev_list, ib_dev_list) { > - if (dev->sys_image_guid == mpi->sys_image_guid) > + if (dev->sys_image_guid == mpi->sys_image_guid && > + mlx5_core_same_coredev_type(dev->mdev, mpi->mdev)) > bound = mlx5_ib_bind_slave_port(dev, mpi); > > if (bound) { > diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h > index fc7e6153b73d..4f9e6f6dbaab 100644 > --- a/include/linux/mlx5/driver.h > +++ b/include/linux/mlx5/driver.h > @@ -1202,6 +1202,12 @@ static inline bool mlx5_core_is_vf(const struct mlx5_core_dev *dev) > return dev->coredev_type == MLX5_COREDEV_VF; > } > > +static inline bool mlx5_core_same_coredev_type(const struct mlx5_core_dev *dev1, > + const struct mlx5_core_dev *dev2) > +{ > + return dev1->coredev_type == dev2->coredev_type; > +} > + > static inline bool mlx5_core_is_ecpf(const struct mlx5_core_dev *dev) > { > return dev->caps.embedded_cpu; Hmmm.. Looks good to me Reviewed-by: Mateusz Polchlopek <mateusz.polchlopek@intel.com>
On Tue, 03 Dec 2024 15:45:37 +0200, Leon Romanovsky wrote: > Different core device types such as PFs and VFs shouldn't be affiliated > together since they have different capabilities, fix that by enforcing > type check before doing the affiliation. > > Applied, thanks! [1/1] RDMA/mlx5: Enforce same type port association for multiport RoCE https://git.kernel.org/rdma/rdma/c/e05feab22fd7da Best regards,
diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c index bc7930d0c564..c2314797afc9 100644 --- a/drivers/infiniband/hw/mlx5/main.c +++ b/drivers/infiniband/hw/mlx5/main.c @@ -3639,7 +3639,8 @@ static int mlx5_ib_init_multiport_master(struct mlx5_ib_dev *dev) list_for_each_entry(mpi, &mlx5_ib_unaffiliated_port_list, list) { if (dev->sys_image_guid == mpi->sys_image_guid && - (mlx5_core_native_port_num(mpi->mdev) - 1) == i) { + (mlx5_core_native_port_num(mpi->mdev) - 1) == i && + mlx5_core_same_coredev_type(dev->mdev, mpi->mdev)) { bound = mlx5_ib_bind_slave_port(dev, mpi); } @@ -4785,7 +4786,8 @@ static int mlx5r_mp_probe(struct auxiliary_device *adev, mutex_lock(&mlx5_ib_multiport_mutex); list_for_each_entry(dev, &mlx5_ib_dev_list, ib_dev_list) { - if (dev->sys_image_guid == mpi->sys_image_guid) + if (dev->sys_image_guid == mpi->sys_image_guid && + mlx5_core_same_coredev_type(dev->mdev, mpi->mdev)) bound = mlx5_ib_bind_slave_port(dev, mpi); if (bound) { diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index fc7e6153b73d..4f9e6f6dbaab 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -1202,6 +1202,12 @@ static inline bool mlx5_core_is_vf(const struct mlx5_core_dev *dev) return dev->coredev_type == MLX5_COREDEV_VF; } +static inline bool mlx5_core_same_coredev_type(const struct mlx5_core_dev *dev1, + const struct mlx5_core_dev *dev2) +{ + return dev1->coredev_type == dev2->coredev_type; +} + static inline bool mlx5_core_is_ecpf(const struct mlx5_core_dev *dev) { return dev->caps.embedded_cpu;