/**
* iser_sg_to_page_vec - Translates scatterlist entries to physical addresses
* and returns the length of resulting physical address array (may be less than
* the original due to possible compaction).
*
* we build a "page vec" under the assumption that the SG meets the RDMA
* alignment requirements. Other then the first and last SG elements, all
* the "internal" elements can be compacted into a list whose elements are
* dma addresses of physical pages. The code supports also the weird case
* where --few fragments of the same page-- are present in the SG as
* consecutive elements. Also, it handles one entry SG.
*/
static int iser_sg_to_page_vec(struct iser_data_buf *data,
struct iser_page_vec *page_vec,
struct ib_device *ibdev)
{
struct scatterlist *sg = (struct scatterlist *)data->buf;
u64 first_addr, last_addr, page;
int end_aligned;
unsigned int cur_page = 0;
unsigned long total_sz = 0;
int i;
/* compute the offset of first element */
page_vec->offset = (u64) sg[0].offset & ~MASK_4K;
for (i = 0; i < data->dma_nents; i++) {
unsigned int dma_len = ib_sg_dma_len(ibdev, &sg[i]);
total_sz += dma_len;
first_addr = ib_sg_dma_address(ibdev, &sg[i]);
last_addr = first_addr + dma_len;
end_aligned = !(last_addr & ~MASK_4K);
/* continue to collect page fragments till aligned or SG ends */
while (!end_aligned && (i + 1 < data->dma_nents)) {
i++;
dma_len = ib_sg_dma_len(ibdev, &sg[i]);
total_sz += dma_len;
last_addr = ib_sg_dma_address(ibdev, &sg[i]) + dma_len;
end_aligned = !(last_addr & ~MASK_4K);
}
/* handle the 1st page in the 1st DMA element */
if (cur_page == 0) {
page = first_addr & MASK_4K;
page_vec->pages[cur_page] = page;
cur_page++;
page += SIZE_4K;
} else
page = first_addr;
for (; page < last_addr; page += SIZE_4K) {
page_vec->pages[cur_page] = page;
cur_page++;
}
}
page_vec->data_size = total_sz;
iser_dbg("page_vec->data_size:%d cur_page %d\n", page_vec->data_size,cur_page);
return cur_page;
}
static int
iser_reg_dma(struct iser_device *device, struct iser_data_buf *mem,
struct iser_mem_reg *reg)
{
struct scatterlist *sg = mem->sg;
reg->sge.lkey = device->pd->local_dma_lkey;
/*
* FIXME: rework the registration code path to differentiate
* rkey/lkey use cases
*/
if (device->pd->flags & IB_PD_UNSAFE_GLOBAL_RKEY)
reg->rkey = device->pd->unsafe_global_rkey;
else
reg->rkey = 0;
reg->sge.addr = ib_sg_dma_address(device->ib_device, &sg[0]);
reg->sge.length = ib_sg_dma_len(device->ib_device, &sg[0]);
iser_dbg("Single DMA entry: lkey=0x%x, rkey=0x%x, addr=0x%llx,"
" length=0x%x\n", reg->sge.lkey, reg->rkey,
reg->sge.addr, reg->sge.length);
return 0;
}
static int rdma_rw_init_single_wr(struct rdma_rw_ctx *ctx, struct ib_qp *qp,
struct scatterlist *sg, u32 offset, u64 remote_addr, u32 rkey,
enum dma_data_direction dir)
{
struct ib_device *dev = qp->pd->device;
struct ib_rdma_wr *rdma_wr = &ctx->single.wr;
ctx->nr_ops = 1;
ctx->single.sge.lkey = qp->pd->local_dma_lkey;
ctx->single.sge.addr = ib_sg_dma_address(dev, sg) + offset;
ctx->single.sge.length = ib_sg_dma_len(dev, sg) - offset;
memset(rdma_wr, 0, sizeof(*rdma_wr));
if (dir == DMA_TO_DEVICE)
rdma_wr->wr.opcode = IB_WR_RDMA_WRITE;
else
rdma_wr->wr.opcode = IB_WR_RDMA_READ;
rdma_wr->wr.sg_list = &ctx->single.sge;
rdma_wr->wr.num_sge = 1;
rdma_wr->remote_addr = remote_addr;
rdma_wr->rkey = rkey;
ctx->type = RDMA_RW_SINGLE_WR;
return 1;
}
static int rdma_rw_init_map_wrs(struct rdma_rw_ctx *ctx, struct ib_qp *qp,
struct scatterlist *sg, u32 sg_cnt, u32 offset,
u64 remote_addr, u32 rkey, enum dma_data_direction dir)
{
struct ib_device *dev = qp->pd->device;
u32 max_sge = rdma_rw_max_sge(dev, dir);
struct ib_sge *sge;
u32 total_len = 0, i, j;
ctx->nr_ops = DIV_ROUND_UP(sg_cnt, max_sge);
ctx->map.sges = sge = kcalloc(sg_cnt, sizeof(*sge), GFP_KERNEL);
if (!ctx->map.sges)
goto out;
ctx->map.wrs = kcalloc(ctx->nr_ops, sizeof(*ctx->map.wrs), GFP_KERNEL);
if (!ctx->map.wrs)
goto out_free_sges;
for (i = 0; i < ctx->nr_ops; i++) {
struct ib_rdma_wr *rdma_wr = &ctx->map.wrs[i];
u32 nr_sge = min(sg_cnt, max_sge);
if (dir == DMA_TO_DEVICE)
rdma_wr->wr.opcode = IB_WR_RDMA_WRITE;
else
rdma_wr->wr.opcode = IB_WR_RDMA_READ;
rdma_wr->remote_addr = remote_addr + total_len;
rdma_wr->rkey = rkey;
rdma_wr->wr.sg_list = sge;
for (j = 0; j < nr_sge; j++, sg = sg_next(sg)) {
rdma_wr->wr.num_sge++;
sge->addr = ib_sg_dma_address(dev, sg) + offset;
sge->length = ib_sg_dma_len(dev, sg) - offset;
sge->lkey = qp->pd->local_dma_lkey;
total_len += sge->length;
sge++;
sg_cnt--;
offset = 0;
}
if (i + 1 < ctx->nr_ops)
rdma_wr->wr.next = &ctx->map.wrs[i + 1].wr;
}
ctx->type = RDMA_RW_MULTI_WR;
return ctx->nr_ops;
out_free_sges:
kfree(ctx->map.sges);
out:
return -ENOMEM;
}
static void iser_data_buf_dump(struct iser_data_buf *data,
struct ib_device *ibdev)
{
struct scatterlist *sg;
int i;
for_each_sg(data->sg, sg, data->dma_nents, i)
iser_dbg("sg[%d] dma_addr:0x%lX page:0x%p "
"off:0x%x sz:0x%x dma_len:0x%x\n",
i, (unsigned long)ib_sg_dma_address(ibdev, sg),
sg_page(sg), sg->offset,
sg->length, ib_sg_dma_len(ibdev, sg));
}
static void iser_data_buf_dump(struct iser_data_buf *data,
struct ib_device *ibdev)
{
struct scatterlist *sg = (struct scatterlist *)data->buf;
int i;
for (i = 0; i < data->dma_nents; i++)
iser_err("sg[%d] dma_addr:0x%lX page:0x%p "
"off:0x%x sz:0x%x dma_len:0x%x\n",
i, (unsigned long)ib_sg_dma_address(ibdev, &sg[i]),
sg[i].page, sg[i].offset,
sg[i].length, ib_sg_dma_len(ibdev, &sg[i]));
}
/**
* iser_data_buf_aligned_len - Tries to determine the maximal correctly aligned
* for RDMA sub-list of a scatter-gather list of memory buffers, and returns
* the number of entries which are aligned correctly. Supports the case where
* consecutive SG elements are actually fragments of the same physcial page.
*/
static unsigned int iser_data_buf_aligned_len(struct iser_data_buf *data,
struct ib_device *ibdev)
{
struct scatterlist *sg;
u64 end_addr, next_addr;
int i, cnt;
unsigned int ret_len = 0;
sg = (struct scatterlist *)data->buf;
for (cnt = 0, i = 0; i < data->dma_nents; i++, cnt++) {
/* iser_dbg("Checking sg iobuf [%d]: phys=0x%08lX "
"offset: %ld sz: %ld\n", i,
(unsigned long)page_to_phys(sg[i].page),
(unsigned long)sg[i].offset,
(unsigned long)sg[i].length); */
end_addr = ib_sg_dma_address(ibdev, &sg[i]) +
ib_sg_dma_len(ibdev, &sg[i]);
/* iser_dbg("Checking sg iobuf end address "
"0x%08lX\n", end_addr); */
if (i + 1 < data->dma_nents) {
next_addr = ib_sg_dma_address(ibdev, &sg[i+1]);
/* are i, i+1 fragments of the same page? */
if (end_addr == next_addr)
continue;
else if (!IS_4K_ALIGNED(end_addr)) {
ret_len = cnt + 1;
break;
}
}
}
if (i == data->dma_nents)
ret_len = cnt; /* loop ended */
iser_dbg("Found %d aligned entries out of %d in sg:0x%p\n",
ret_len, data->dma_nents, data);
return ret_len;
}
int rds_ib_map_fmr(struct rds_ib_device *rds_ibdev, struct rds_ib_mr *ibmr,
struct scatterlist *sg, unsigned int nents)
{
struct ib_device *dev = rds_ibdev->dev;
struct rds_ib_fmr *fmr = &ibmr->u.fmr;
struct scatterlist *scat = sg;
u64 io_addr = 0;
u64 *dma_pages;
u32 len;
int page_cnt, sg_dma_len;
int i, j;
int ret;
sg_dma_len = ib_dma_map_sg(dev, sg, nents, DMA_BIDIRECTIONAL);
if (unlikely(!sg_dma_len)) {
pr_warn("RDS/IB: %s failed!\n", __func__);
return -EBUSY;
}
len = 0;
page_cnt = 0;
for (i = 0; i < sg_dma_len; ++i) {
unsigned int dma_len = ib_sg_dma_len(dev, &scat[i]);
u64 dma_addr = ib_sg_dma_address(dev, &scat[i]);
if (dma_addr & ~PAGE_MASK) {
if (i > 0)
return -EINVAL;
else
++page_cnt;
}
if ((dma_addr + dma_len) & ~PAGE_MASK) {
if (i < sg_dma_len - 1)
return -EINVAL;
else
++page_cnt;
}
len += dma_len;
}
page_cnt += len >> PAGE_SHIFT;
if (page_cnt > ibmr->pool->fmr_attr.max_pages)
return -EINVAL;
dma_pages = kmalloc_node(sizeof(u64) * page_cnt, GFP_ATOMIC,
rdsibdev_to_node(rds_ibdev));
if (!dma_pages)
return -ENOMEM;
page_cnt = 0;
for (i = 0; i < sg_dma_len; ++i) {
unsigned int dma_len = ib_sg_dma_len(dev, &scat[i]);
u64 dma_addr = ib_sg_dma_address(dev, &scat[i]);
for (j = 0; j < dma_len; j += PAGE_SIZE)
dma_pages[page_cnt++] =
(dma_addr & PAGE_MASK) + j;
}
ret = ib_map_phys_fmr(fmr->fmr, dma_pages, page_cnt, io_addr);
if (ret)
goto out;
/* Success - we successfully remapped the MR, so we can
* safely tear down the old mapping.
*/
rds_ib_teardown_mr(ibmr);
ibmr->sg = scat;
ibmr->sg_len = nents;
ibmr->sg_dma_len = sg_dma_len;
ibmr->remap_count++;
if (ibmr->pool->pool_type == RDS_IB_MR_8K_POOL)
rds_ib_stats_inc(s_ib_rdma_mr_8k_used);
else
rds_ib_stats_inc(s_ib_rdma_mr_1m_used);
ret = 0;
out:
kfree(dma_pages);
return ret;
}
static int rds_ib_map_frmr(struct rds_ib_device *rds_ibdev,
struct rds_ib_mr_pool *pool,
struct rds_ib_mr *ibmr,
struct scatterlist *sg, unsigned int sg_len)
{
struct ib_device *dev = rds_ibdev->dev;
struct rds_ib_frmr *frmr = &ibmr->u.frmr;
int i;
u32 len;
int ret = 0;
/* We want to teardown old ibmr values here and fill it up with
* new sg values
*/
rds_ib_teardown_mr(ibmr);
ibmr->sg = sg;
ibmr->sg_len = sg_len;
ibmr->sg_dma_len = 0;
frmr->sg_byte_len = 0;
WARN_ON(ibmr->sg_dma_len);
ibmr->sg_dma_len = ib_dma_map_sg(dev, ibmr->sg, ibmr->sg_len,
DMA_BIDIRECTIONAL);
if (unlikely(!ibmr->sg_dma_len)) {
pr_warn("RDS/IB: %s failed!\n", __func__);
return -EBUSY;
}
frmr->sg_byte_len = 0;
frmr->dma_npages = 0;
len = 0;
ret = -EINVAL;
for (i = 0; i < ibmr->sg_dma_len; ++i) {
unsigned int dma_len = ib_sg_dma_len(dev, &ibmr->sg[i]);
u64 dma_addr = ib_sg_dma_address(dev, &ibmr->sg[i]);
frmr->sg_byte_len += dma_len;
if (dma_addr & ~PAGE_MASK) {
if (i > 0)
goto out_unmap;
else
++frmr->dma_npages;
}
if ((dma_addr + dma_len) & ~PAGE_MASK) {
if (i < ibmr->sg_dma_len - 1)
goto out_unmap;
else
++frmr->dma_npages;
}
len += dma_len;
}
frmr->dma_npages += len >> PAGE_SHIFT;
if (frmr->dma_npages > ibmr->pool->fmr_attr.max_pages) {
ret = -EMSGSIZE;
goto out_unmap;
}
ret = rds_ib_post_reg_frmr(ibmr);
if (ret)
goto out_unmap;
if (ibmr->pool->pool_type == RDS_IB_MR_8K_POOL)
rds_ib_stats_inc(s_ib_rdma_mr_8k_used);
else
rds_ib_stats_inc(s_ib_rdma_mr_1m_used);
return ret;
out_unmap:
ib_dma_unmap_sg(rds_ibdev->dev, ibmr->sg, ibmr->sg_len,
DMA_BIDIRECTIONAL);
ibmr->sg_dma_len = 0;
return ret;
}
/**
* iser_reg_rdma_mem - Registers memory intended for RDMA,
* obtaining rkey and va
*
* returns 0 on success, errno code on failure
*/
int iser_reg_rdma_mem(struct iscsi_iser_cmd_task *iser_ctask,
enum iser_data_dir cmd_dir)
{
struct iser_conn *ib_conn = iser_ctask->iser_conn->ib_conn;
struct iser_device *device = ib_conn->device;
struct ib_device *ibdev = device->ib_device;
struct iser_data_buf *mem = &iser_ctask->data[cmd_dir];
struct iser_regd_buf *regd_buf;
int aligned_len;
int err;
int i;
struct scatterlist *sg;
regd_buf = &iser_ctask->rdma_regd[cmd_dir];
aligned_len = iser_data_buf_aligned_len(mem, ibdev);
if (aligned_len != mem->dma_nents) {
iser_err("rdma alignment violation %d/%d aligned\n",
aligned_len, mem->size);
iser_data_buf_dump(mem, ibdev);
/* unmap the command data before accessing it */
iser_dma_unmap_task_data(iser_ctask);
/* allocate copy buf, if we are writing, copy the */
/* unaligned scatterlist, dma map the copy */
if (iser_start_rdma_unaligned_sg(iser_ctask, cmd_dir) != 0)
return -ENOMEM;
mem = &iser_ctask->data_copy[cmd_dir];
}
/* if there a single dma entry, FMR is not needed */
if (mem->dma_nents == 1) {
sg = (struct scatterlist *)mem->buf;
regd_buf->reg.lkey = device->mr->lkey;
regd_buf->reg.rkey = device->mr->rkey;
regd_buf->reg.len = ib_sg_dma_len(ibdev, &sg[0]);
regd_buf->reg.va = ib_sg_dma_address(ibdev, &sg[0]);
regd_buf->reg.is_fmr = 0;
iser_dbg("PHYSICAL Mem.register: lkey: 0x%08X rkey: 0x%08X "
"va: 0x%08lX sz: %ld]\n",
(unsigned int)regd_buf->reg.lkey,
(unsigned int)regd_buf->reg.rkey,
(unsigned long)regd_buf->reg.va,
(unsigned long)regd_buf->reg.len);
} else { /* use FMR for multiple dma entries */
iser_page_vec_build(mem, ib_conn->page_vec, ibdev);
err = iser_reg_page_vec(ib_conn, ib_conn->page_vec, ®d_buf->reg);
if (err) {
iser_data_buf_dump(mem, ibdev);
iser_err("mem->dma_nents = %d (dlength = 0x%x)\n", mem->dma_nents,
ntoh24(iser_ctask->desc.iscsi_header.dlength));
iser_err("page_vec: data_size = 0x%x, length = %d, offset = 0x%x\n",
ib_conn->page_vec->data_size, ib_conn->page_vec->length,
ib_conn->page_vec->offset);
for (i=0 ; i<ib_conn->page_vec->length ; i++)
iser_err("page_vec[%d] = 0x%llx\n", i,
(unsigned long long) ib_conn->page_vec->pages[i]);
return err;
}
}
/* take a reference on this regd buf such that it will not be released *
* (eg in send dto completion) before we get the scsi response */
atomic_inc(®d_buf->ref_count);
return 0;
}
