IB: split struct ib_send_wr
This patch split up struct ib_send_wr so that all non-trivial verbs
use their own structure which embedds struct ib_send_wr. This dramaticly
shrinks the size of a WR for most common operations:
sizeof(struct ib_send_wr) (old): 96
sizeof(struct ib_send_wr): 48
sizeof(struct ib_rdma_wr): 64
sizeof(struct ib_atomic_wr): 96
sizeof(struct ib_ud_wr): 88
sizeof(struct ib_fast_reg_wr): 88
sizeof(struct ib_bind_mw_wr): 96
sizeof(struct ib_sig_handover_wr): 80
And with Sagi's pending MR rework the fast registration WR will also be
down to a reasonable size:
sizeof(struct ib_fastreg_wr): 64
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Bart Van Assche <bart.vanassche@sandisk.com> [srp, srpt]
Reviewed-by: Chuck Lever <chuck.lever@oracle.com> [sunrpc]
Tested-by: Haggai Eran <haggaie@mellanox.com>
Tested-by: Sagi Grimberg <sagig@mellanox.com>
Tested-by: Steve Wise <swise@opengridcomputing.com>
diff --git a/drivers/infiniband/hw/nes/nes_verbs.c b/drivers/infiniband/hw/nes/nes_verbs.c
index 44cb513..f71b37b 100644
--- a/drivers/infiniband/hw/nes/nes_verbs.c
+++ b/drivers/infiniband/hw/nes/nes_verbs.c
@@ -3372,9 +3372,9 @@
wqe_misc |= NES_IWARP_SQ_WQE_LOCAL_FENCE;
set_wqe_32bit_value(wqe->wqe_words, NES_IWARP_SQ_WQE_RDMA_STAG_IDX,
- ib_wr->wr.rdma.rkey);
+ rdma_wr(ib_wr)->rkey);
set_wqe_64bit_value(wqe->wqe_words, NES_IWARP_SQ_WQE_RDMA_TO_LOW_IDX,
- ib_wr->wr.rdma.remote_addr);
+ rdma_wr(ib_wr)->remote_addr);
if ((ib_wr->send_flags & IB_SEND_INLINE) &&
((nes_drv_opt & NES_DRV_OPT_NO_INLINE_DATA) == 0) &&
@@ -3409,9 +3409,9 @@
}
set_wqe_64bit_value(wqe->wqe_words, NES_IWARP_SQ_WQE_RDMA_TO_LOW_IDX,
- ib_wr->wr.rdma.remote_addr);
+ rdma_wr(ib_wr)->remote_addr);
set_wqe_32bit_value(wqe->wqe_words, NES_IWARP_SQ_WQE_RDMA_STAG_IDX,
- ib_wr->wr.rdma.rkey);
+ rdma_wr(ib_wr)->rkey);
set_wqe_32bit_value(wqe->wqe_words, NES_IWARP_SQ_WQE_RDMA_LENGTH_IDX,
ib_wr->sg_list->length);
set_wqe_64bit_value(wqe->wqe_words, NES_IWARP_SQ_WQE_FRAG0_LOW_IDX,
@@ -3428,15 +3428,16 @@
case IB_WR_FAST_REG_MR:
{
int i;
- int flags = ib_wr->wr.fast_reg.access_flags;
+ struct ib_fast_reg_wr *fwr = fast_reg_wr(ib_wr);
+ int flags = fwr->access_flags;
struct nes_ib_fast_reg_page_list *pnesfrpl =
- container_of(ib_wr->wr.fast_reg.page_list,
+ container_of(fwr->page_list,
struct nes_ib_fast_reg_page_list,
ibfrpl);
u64 *src_page_list = pnesfrpl->ibfrpl.page_list;
u64 *dst_page_list = pnesfrpl->nes_wqe_pbl.kva;
- if (ib_wr->wr.fast_reg.page_list_len >
+ if (fwr->page_list_len >
(NES_4K_PBL_CHUNK_SIZE / sizeof(u64))) {
nes_debug(NES_DBG_IW_TX, "SQ_FMR: bad page_list_len\n");
err = -EINVAL;
@@ -3445,19 +3446,19 @@
wqe_misc = NES_IWARP_SQ_OP_FAST_REG;
set_wqe_64bit_value(wqe->wqe_words,
NES_IWARP_SQ_FMR_WQE_VA_FBO_LOW_IDX,
- ib_wr->wr.fast_reg.iova_start);
+ fwr->iova_start);
set_wqe_32bit_value(wqe->wqe_words,
NES_IWARP_SQ_FMR_WQE_LENGTH_LOW_IDX,
- ib_wr->wr.fast_reg.length);
+ fwr->length);
set_wqe_32bit_value(wqe->wqe_words,
NES_IWARP_SQ_FMR_WQE_LENGTH_HIGH_IDX, 0);
set_wqe_32bit_value(wqe->wqe_words,
NES_IWARP_SQ_FMR_WQE_MR_STAG_IDX,
- ib_wr->wr.fast_reg.rkey);
+ fwr->rkey);
/* Set page size: */
- if (ib_wr->wr.fast_reg.page_shift == 12) {
+ if (fwr->page_shift == 12) {
wqe_misc |= NES_IWARP_SQ_FMR_WQE_PAGE_SIZE_4K;
- } else if (ib_wr->wr.fast_reg.page_shift == 21) {
+ } else if (fwr->page_shift == 21) {
wqe_misc |= NES_IWARP_SQ_FMR_WQE_PAGE_SIZE_2M;
} else {
nes_debug(NES_DBG_IW_TX, "Invalid page shift,"
@@ -3480,11 +3481,11 @@
wqe_misc |= NES_IWARP_SQ_FMR_WQE_RIGHTS_ENABLE_WINDOW_BIND;
/* Fill in PBL info: */
- if (ib_wr->wr.fast_reg.page_list_len >
+ if (fwr->page_list_len >
pnesfrpl->ibfrpl.max_page_list_len) {
nes_debug(NES_DBG_IW_TX, "Invalid page list length,"
" ib_wr=%p, value=%u, max=%u\n",
- ib_wr, ib_wr->wr.fast_reg.page_list_len,
+ ib_wr, fwr->page_list_len,
pnesfrpl->ibfrpl.max_page_list_len);
err = -EINVAL;
break;
@@ -3496,19 +3497,19 @@
set_wqe_32bit_value(wqe->wqe_words,
NES_IWARP_SQ_FMR_WQE_PBL_LENGTH_IDX,
- ib_wr->wr.fast_reg.page_list_len * 8);
+ fwr->page_list_len * 8);
- for (i = 0; i < ib_wr->wr.fast_reg.page_list_len; i++)
+ for (i = 0; i < fwr->page_list_len; i++)
dst_page_list[i] = cpu_to_le64(src_page_list[i]);
nes_debug(NES_DBG_IW_TX, "SQ_FMR: iova_start: %llx, "
"length: %d, rkey: %0x, pgl_paddr: %llx, "
"page_list_len: %u, wqe_misc: %x\n",
- (unsigned long long) ib_wr->wr.fast_reg.iova_start,
- ib_wr->wr.fast_reg.length,
- ib_wr->wr.fast_reg.rkey,
+ (unsigned long long) fwr->iova_start,
+ fwr->length,
+ fwr->rkey,
(unsigned long long) pnesfrpl->nes_wqe_pbl.paddr,
- ib_wr->wr.fast_reg.page_list_len,
+ fwr->page_list_len,
wqe_misc);
break;
}