xiaomi-sm8450/android_kernel_xiaomi_sm8450
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

net/mlx5: Use order-0 allocations for all WQ types

Browse Source 
Complete the transition of all WQ types to use fragmented
order-0 coherent memory instead of high-order allocations.

CQ-WQ already uses order-0.
Here we do the same for cyclic and linked-list WQs.

This allows the driver to load cleanly on systems with a highly
fragmented coherent memory.

Performance tests:
ConnectX-5 100Gbps, CPU: Intel(R) Xeon(R) CPU E5-2680 v3 @ 2.50GHz
Packet rate of 64B packets, single transmit ring, size 8K.

No degradation is sensed.

Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
This commit is contained in:
Tariq Toukan
2018-04-04 12:54:23 +03:00
committed by Saeed Mahameed
parent 549322f2f9
commit 3a2f703312
9 changed files with 123 additions and 94 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
include/linux/mlx5/driver.h
View File

@@ -983,16 +983,24 @@ static inline u32 mlx5_base_mkey(const u32 key)
return key & 0xffffff00u;
}
static inline void mlx5_core_init_cq_frag_buf(struct mlx5_frag_buf_ctrl *fbc,
void *cqc)
static inline void mlx5_fill_fbc(u8 log_stride, u8 log_sz,
struct mlx5_frag_buf_ctrl *fbc)
{
fbc->log_stride = 6 + MLX5_GET(cqc, cqc, cqe_sz);
fbc->log_sz = MLX5_GET(cqc, cqc, log_cq_size);
fbc->log_stride = log_stride;
fbc->log_sz = log_sz;
fbc->sz_m1 = (1 << fbc->log_sz) - 1;
fbc->log_frag_strides = PAGE_SHIFT - fbc->log_stride;
fbc->frag_sz_m1 = (1 << fbc->log_frag_strides) - 1;
}
static inline void mlx5_core_init_cq_frag_buf(struct mlx5_frag_buf_ctrl *fbc,
void *cqc)
{
mlx5_fill_fbc(6 + MLX5_GET(cqc, cqc, cqe_sz),
MLX5_GET(cqc, cqc, log_cq_size),
fbc);
}
static inline void *mlx5_frag_buf_get_wqe(struct mlx5_frag_buf_ctrl *fbc,
u32 ix)
{