RDMA/bnxt_re: Use rdma_umem_for_each_dma_block()

Browse source

This driver is taking the SGL out of the umem and passing it through a
struct bnxt_qplib_sg_info. Instead of passing the SGL pass the umem and
then use rdma_umem_for_each_dma_block() directly.

Move the calls of ib_umem_num_dma_blocks() closer to their actual point of
use, npages is only set for non-umem pbl flows.

Link: https://lore.kernel.org/r/0-v1-b37437a73f35+49c-bnxt_re_dma_block_jgg@nvidia.com
Acked-by: Selvin Xavier <selvin.xavier@broadcom.com>
Tested-by: Selvin Xavier <selvin.xavier@broadcom.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>

...

This commit is contained in:

Jason Gunthorpe 2020-09-29 21:24:35 -03:00

parent 5ce2dced8e

commit 6ef999f500

3 changed files with 22 additions and 29 deletions

Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL

Download patch file Download diff file Expand all files Collapse all files

18

drivers/infiniband/hw/bnxt_re/ib_verbs.c

View file

30

drivers/infiniband/hw/bnxt_re/qplib_res.c

View file

3

drivers/infiniband/hw/bnxt_re/qplib_res.h

View file