diff options
author | Håkon Bugge <haakon.bugge@oracle.com> | 2020-08-03 09:19:38 +0300 |
---|---|---|
committer | Jason Gunthorpe <jgg@nvidia.com> | 2020-08-24 17:31:21 +0300 |
commit | 0ae207fb91a897780f0853864d80c48edec7f374 (patch) | |
tree | 3c1665ef719b2e26a32b8387277507ba472c1fa0 /drivers/infiniband/hw/mlx4/mlx4_ib.h | |
parent | e7d087fce63fada387afcc6c31a2eaf81452baee (diff) | |
download | linux-0ae207fb91a897780f0853864d80c48edec7f374.tar.xz |
IB/mlx4: Separate tunnel and wire bufs parameters
Using CX-3 in virtualized mode, MAD packets are proxied through the PF
driver. The feed is N tunnel QPs, and what is received from the VFs is
multiplexed out on the wire QP. Since this is a many-to-one scenario, it
is better to have separate initialization parameters for the two usages.
The number of wire and tunnel bufs are yanked up to 2K and 512
respectively. With this set of parameters, a system consisting of eight
physical servers, each with eight VMs and 14 I/O servers (BM), can run
switch fail-over without seeing:
mlx4_ib_demux_mad: failed sending GSI to slave 3 via tunnel qp (-11)
or
mlx4_ib_multiplex_mad: failed sending GSI to wire on behalf of slave 2 (-11)
Fixes: 3cf69cc8dbeb ("IB/mlx4: Add CM paravirtualization")
Link: https://lore.kernel.org/r/20200803061941.1139994-4-haakon.bugge@oracle.com
Signed-off-by: Håkon Bugge <haakon.bugge@oracle.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Diffstat (limited to 'drivers/infiniband/hw/mlx4/mlx4_ib.h')
-rw-r--r-- | drivers/infiniband/hw/mlx4/mlx4_ib.h | 3 |
1 files changed, 2 insertions, 1 deletions
diff --git a/drivers/infiniband/hw/mlx4/mlx4_ib.h b/drivers/infiniband/hw/mlx4/mlx4_ib.h index 38e87a700a2a..db3cc61de0db 100644 --- a/drivers/infiniband/hw/mlx4/mlx4_ib.h +++ b/drivers/infiniband/hw/mlx4/mlx4_ib.h @@ -233,7 +233,8 @@ enum mlx4_ib_mad_ifc_flags { }; enum { - MLX4_NUM_TUNNEL_BUFS = 256, + MLX4_NUM_TUNNEL_BUFS = 512, + MLX4_NUM_WIRE_BUFS = 2048, }; struct mlx4_ib_tunnel_header { |