RDMA/rtrs_clt: Alloc less memory with write path fast memory registration

With write path fast memory registration, we need less memory for
each request.

With fast memory registration, we can reduce max_send_sge to save
memory usage.

Also convert the kmalloc_array to kcalloc.

Link: https://lore.kernel.org/r/20210621055340.11789-4-jinpu.wang@ionos.com
Signed-off-by: Jack Wang <jinpu.wang@cloud.ionos.com>
Reviewed-by: Md Haris Iqbal <haris.iqbal@cloud.ionos.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
This commit is contained in:
Jack Wang 2021-06-21 07:53:38 +02:00 committed by Jason Gunthorpe
parent 2ece9ec62e
commit df1885a755

View File

@ -1372,8 +1372,7 @@ static int alloc_sess_reqs(struct rtrs_clt_sess *sess)
if (!req->iu)
goto out;
req->sge = kmalloc_array(clt->max_segments + 1,
sizeof(*req->sge), GFP_KERNEL);
req->sge = kcalloc(2, sizeof(*req->sge), GFP_KERNEL);
if (!req->sge)
goto out;
@ -1675,7 +1674,7 @@ static int create_con_cq_qp(struct rtrs_clt_con *con)
sess->queue_depth * 3 + 1);
max_recv_wr = min_t(int, wr_limit,
sess->queue_depth * 3 + 1);
max_send_sge = sess->clt->max_segments + 1;
max_send_sge = 2;
}
cq_num = max_send_wr + max_recv_wr;
/* alloc iu to recv new rkey reply when server reports flags set */