--- zzzz-none-000/linux-3.10.107/drivers/infiniband/ulp/iser/iser_memory.c 2017-06-27 09:49:32.000000000 +0000 +++ scorpion-7490-727/linux-3.10.107/drivers/infiniband/ulp/iser/iser_memory.c 2021-02-04 17:41:59.000000000 +0000 @@ -1,6 +1,6 @@ /* * Copyright (c) 2004, 2005, 2006 Voltaire, Inc. All rights reserved. - * Copyright (c) 2013 Mellanox Technologies. All rights reserved. + * Copyright (c) 2013-2014 Mellanox Technologies. All rights reserved. * * This software is available to you under a choice of one of two * licenses. You may choose to be licensed under the terms of the GNU @@ -38,120 +38,97 @@ #include #include "iscsi_iser.h" +static +int iser_fast_reg_fmr(struct iscsi_iser_task *iser_task, + struct iser_data_buf *mem, + struct iser_reg_resources *rsc, + struct iser_mem_reg *mem_reg); +static +int iser_fast_reg_mr(struct iscsi_iser_task *iser_task, + struct iser_data_buf *mem, + struct iser_reg_resources *rsc, + struct iser_mem_reg *mem_reg); + +static struct iser_reg_ops fastreg_ops = { + .alloc_reg_res = iser_alloc_fastreg_pool, + .free_reg_res = iser_free_fastreg_pool, + .reg_mem = iser_fast_reg_mr, + .unreg_mem = iser_unreg_mem_fastreg, + .reg_desc_get = iser_reg_desc_get_fr, + .reg_desc_put = iser_reg_desc_put_fr, +}; + +static struct iser_reg_ops fmr_ops = { + .alloc_reg_res = iser_alloc_fmr_pool, + .free_reg_res = iser_free_fmr_pool, + .reg_mem = iser_fast_reg_fmr, + .unreg_mem = iser_unreg_mem_fmr, + .reg_desc_get = iser_reg_desc_get_fmr, + .reg_desc_put = iser_reg_desc_put_fmr, +}; -#define ISER_KMALLOC_THRESHOLD 0x20000 /* 128K - kmalloc limit */ - -/** - * iser_start_rdma_unaligned_sg - */ -static int iser_start_rdma_unaligned_sg(struct iscsi_iser_task *iser_task, - enum iser_data_dir cmd_dir) +int iser_assign_reg_ops(struct iser_device *device) { - int dma_nents; - struct ib_device *dev; - char *mem = NULL; - struct iser_data_buf *data = &iser_task->data[cmd_dir]; - unsigned long cmd_data_len = data->data_len; - - if (cmd_data_len > ISER_KMALLOC_THRESHOLD) - mem = (void *)__get_free_pages(GFP_ATOMIC, - ilog2(roundup_pow_of_two(cmd_data_len)) - PAGE_SHIFT); - else - mem = kmalloc(cmd_data_len, GFP_ATOMIC); + struct ib_device_attr *dev_attr = &device->dev_attr; - if (mem == NULL) { - iser_err("Failed to allocate mem size %d %d for copying sglist\n", - data->size,(int)cmd_data_len); - return -ENOMEM; + /* Assign function handles - based on FMR support */ + if (device->ib_device->alloc_fmr && device->ib_device->dealloc_fmr && + device->ib_device->map_phys_fmr && device->ib_device->unmap_fmr) { + iser_info("FMR supported, using FMR for registration\n"); + device->reg_ops = &fmr_ops; + } else + if (dev_attr->device_cap_flags & IB_DEVICE_MEM_MGT_EXTENSIONS) { + iser_info("FastReg supported, using FastReg for registration\n"); + device->reg_ops = &fastreg_ops; + } else { + iser_err("IB device does not support FMRs nor FastRegs, can't register memory\n"); + return -1; } - if (cmd_dir == ISER_DIR_OUT) { - /* copy the unaligned sg the buffer which is used for RDMA */ - struct scatterlist *sgl = (struct scatterlist *)data->buf; - struct scatterlist *sg; - int i; - char *p, *from; - - p = mem; - for_each_sg(sgl, sg, data->size, i) { - from = kmap_atomic(sg_page(sg)); - memcpy(p, - from + sg->offset, - sg->length); - kunmap_atomic(from); - p += sg->length; - } - } - - sg_init_one(&iser_task->data_copy[cmd_dir].sg_single, mem, cmd_data_len); - iser_task->data_copy[cmd_dir].buf = - &iser_task->data_copy[cmd_dir].sg_single; - iser_task->data_copy[cmd_dir].size = 1; - - iser_task->data_copy[cmd_dir].copy_buf = mem; - - dev = iser_task->iser_conn->ib_conn->device->ib_device; - dma_nents = ib_dma_map_sg(dev, - &iser_task->data_copy[cmd_dir].sg_single, - 1, - (cmd_dir == ISER_DIR_OUT) ? - DMA_TO_DEVICE : DMA_FROM_DEVICE); - BUG_ON(dma_nents == 0); - - iser_task->data_copy[cmd_dir].dma_nents = dma_nents; return 0; } -/** - * iser_finalize_rdma_unaligned_sg - */ -void iser_finalize_rdma_unaligned_sg(struct iscsi_iser_task *iser_task, - enum iser_data_dir cmd_dir) +struct iser_fr_desc * +iser_reg_desc_get_fr(struct ib_conn *ib_conn) { - struct ib_device *dev; - struct iser_data_buf *mem_copy; - unsigned long cmd_data_len; + struct iser_fr_pool *fr_pool = &ib_conn->fr_pool; + struct iser_fr_desc *desc; + unsigned long flags; + + spin_lock_irqsave(&fr_pool->lock, flags); + desc = list_first_entry(&fr_pool->list, + struct iser_fr_desc, list); + list_del(&desc->list); + spin_unlock_irqrestore(&fr_pool->lock, flags); - dev = iser_task->iser_conn->ib_conn->device->ib_device; - mem_copy = &iser_task->data_copy[cmd_dir]; + return desc; +} - ib_dma_unmap_sg(dev, &mem_copy->sg_single, 1, - (cmd_dir == ISER_DIR_OUT) ? - DMA_TO_DEVICE : DMA_FROM_DEVICE); - - if (cmd_dir == ISER_DIR_IN) { - char *mem; - struct scatterlist *sgl, *sg; - unsigned char *p, *to; - unsigned int sg_size; - int i; - - /* copy back read RDMA to unaligned sg */ - mem = mem_copy->copy_buf; - - sgl = (struct scatterlist *)iser_task->data[ISER_DIR_IN].buf; - sg_size = iser_task->data[ISER_DIR_IN].size; - - p = mem; - for_each_sg(sgl, sg, sg_size, i) { - to = kmap_atomic(sg_page(sg)); - memcpy(to + sg->offset, - p, - sg->length); - kunmap_atomic(to); - p += sg->length; - } - } +void +iser_reg_desc_put_fr(struct ib_conn *ib_conn, + struct iser_fr_desc *desc) +{ + struct iser_fr_pool *fr_pool = &ib_conn->fr_pool; + unsigned long flags; - cmd_data_len = iser_task->data[cmd_dir].data_len; + spin_lock_irqsave(&fr_pool->lock, flags); + list_add(&desc->list, &fr_pool->list); + spin_unlock_irqrestore(&fr_pool->lock, flags); +} - if (cmd_data_len > ISER_KMALLOC_THRESHOLD) - free_pages((unsigned long)mem_copy->copy_buf, - ilog2(roundup_pow_of_two(cmd_data_len)) - PAGE_SHIFT); - else - kfree(mem_copy->copy_buf); +struct iser_fr_desc * +iser_reg_desc_get_fmr(struct ib_conn *ib_conn) +{ + struct iser_fr_pool *fr_pool = &ib_conn->fr_pool; + + return list_first_entry(&fr_pool->list, + struct iser_fr_desc, list); +} - mem_copy->copy_buf = NULL; +void +iser_reg_desc_put_fmr(struct ib_conn *ib_conn, + struct iser_fr_desc *desc) +{ } #define IS_4K_ALIGNED(addr) ((((unsigned long)addr) & ~MASK_4K) == 0) @@ -170,17 +147,17 @@ */ static int iser_sg_to_page_vec(struct iser_data_buf *data, - struct iser_page_vec *page_vec, - struct ib_device *ibdev) + struct ib_device *ibdev, u64 *pages, + int *offset, int *data_size) { - struct scatterlist *sg, *sgl = (struct scatterlist *)data->buf; + struct scatterlist *sg, *sgl = data->sg; u64 start_addr, end_addr, page, chunk_start = 0; unsigned long total_sz = 0; unsigned int dma_len; int i, new_chunk, cur_page, last_ent = data->dma_nents - 1; /* compute the offset of first element */ - page_vec->offset = (u64) sgl[0].offset & ~MASK_4K; + *offset = (u64) sgl[0].offset & ~MASK_4K; new_chunk = 1; cur_page = 0; @@ -204,74 +181,25 @@ which might be unaligned */ page = chunk_start & MASK_4K; do { - page_vec->pages[cur_page++] = page; + pages[cur_page++] = page; page += SIZE_4K; } while (page < end_addr); } - page_vec->data_size = total_sz; - iser_dbg("page_vec->data_size:%d cur_page %d\n", page_vec->data_size,cur_page); + *data_size = total_sz; + iser_dbg("page_vec->data_size:%d cur_page %d\n", + *data_size, cur_page); return cur_page; } - -/** - * 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 int iser_data_buf_aligned_len(struct iser_data_buf *data, - struct ib_device *ibdev) -{ - struct scatterlist *sgl, *sg, *next_sg = NULL; - u64 start_addr, end_addr; - int i, ret_len, start_check = 0; - - if (data->dma_nents == 1) - return 1; - - sgl = (struct scatterlist *)data->buf; - start_addr = ib_sg_dma_address(ibdev, sgl); - - for_each_sg(sgl, sg, data->dma_nents, i) { - if (start_check && !IS_4K_ALIGNED(start_addr)) - break; - - next_sg = sg_next(sg); - if (!next_sg) - break; - - end_addr = start_addr + ib_sg_dma_len(ibdev, sg); - start_addr = ib_sg_dma_address(ibdev, next_sg); - - if (end_addr == start_addr) { - start_check = 0; - continue; - } else - start_check = 1; - - if (!IS_4K_ALIGNED(end_addr)) - break; - } - ret_len = (next_sg) ? i : i+1; - iser_dbg("Found %d aligned entries out of %d in sg:0x%p\n", - ret_len, data->dma_nents, data); - return ret_len; -} - static void iser_data_buf_dump(struct iser_data_buf *data, struct ib_device *ibdev) { - struct scatterlist *sgl = (struct scatterlist *)data->buf; struct scatterlist *sg; int i; - if (iser_debug_level == 0) - return; - - for_each_sg(sgl, sg, data->dma_nents, i) - iser_warn("sg[%d] dma_addr:0x%lX page:0x%p " + 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, @@ -288,29 +216,6 @@ iser_err("%d %lx\n",i,(unsigned long)page_vec->pages[i]); } -static void iser_page_vec_build(struct iser_data_buf *data, - struct iser_page_vec *page_vec, - struct ib_device *ibdev) -{ - int page_vec_len = 0; - - page_vec->length = 0; - page_vec->offset = 0; - - iser_dbg("Translating sg sz: %d\n", data->dma_nents); - page_vec_len = iser_sg_to_page_vec(data, page_vec, ibdev); - iser_dbg("sg len %d page_vec_len %d\n", data->dma_nents,page_vec_len); - - page_vec->length = page_vec_len; - - if (page_vec_len * SIZE_4K < page_vec->data_size) { - iser_err("page_vec too short to hold this SG\n"); - iser_data_buf_dump(data, ibdev); - iser_dump_page_vec(page_vec); - BUG(); - } -} - int iser_dma_map_task_data(struct iscsi_iser_task *iser_task, struct iser_data_buf *data, enum iser_data_dir iser_dir, @@ -319,9 +224,9 @@ struct ib_device *dev; iser_task->dir[iser_dir] = 1; - dev = iser_task->iser_conn->ib_conn->device->ib_device; + dev = iser_task->iser_conn->ib_conn.device->ib_device; - data->dma_nents = ib_dma_map_sg(dev, data->buf, data->size, dma_dir); + data->dma_nents = ib_dma_map_sg(dev, data->sg, data->size, dma_dir); if (data->dma_nents == 0) { iser_err("dma_map_sg failed!!!\n"); return -EINVAL; @@ -329,97 +234,386 @@ return 0; } -void iser_dma_unmap_task_data(struct iscsi_iser_task *iser_task) +void iser_dma_unmap_task_data(struct iscsi_iser_task *iser_task, + struct iser_data_buf *data, + enum dma_data_direction dir) { struct ib_device *dev; - struct iser_data_buf *data; - dev = iser_task->iser_conn->ib_conn->device->ib_device; + dev = iser_task->iser_conn->ib_conn.device->ib_device; + ib_dma_unmap_sg(dev, data->sg, data->size, dir); +} - if (iser_task->dir[ISER_DIR_IN]) { - data = &iser_task->data[ISER_DIR_IN]; - ib_dma_unmap_sg(dev, data->buf, data->size, DMA_FROM_DEVICE); - } +static int +iser_reg_dma(struct iser_device *device, struct iser_data_buf *mem, + struct iser_mem_reg *reg) +{ + struct scatterlist *sg = mem->sg; - if (iser_task->dir[ISER_DIR_OUT]) { - data = &iser_task->data[ISER_DIR_OUT]; - ib_dma_unmap_sg(dev, data->buf, data->size, DMA_TO_DEVICE); - } + reg->sge.lkey = device->pd->local_dma_lkey; + reg->rkey = device->mr->rkey; + 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; } /** - * iser_reg_rdma_mem - Registers memory intended for RDMA, - * obtaining rkey and va + * iser_reg_page_vec - Register physical memory * - * returns 0 on success, errno code on failure + * returns: 0 on success, errno code on failure */ -int iser_reg_rdma_mem(struct iscsi_iser_task *iser_task, - enum iser_data_dir cmd_dir) +static +int iser_fast_reg_fmr(struct iscsi_iser_task *iser_task, + struct iser_data_buf *mem, + struct iser_reg_resources *rsc, + struct iser_mem_reg *reg) { - struct iscsi_conn *iscsi_conn = iser_task->iser_conn->iscsi_conn; - struct iser_conn *ib_conn = iser_task->iser_conn->ib_conn; - struct iser_device *device = ib_conn->device; - struct ib_device *ibdev = device->ib_device; - struct iser_data_buf *mem = &iser_task->data[cmd_dir]; - struct iser_regd_buf *regd_buf; - int aligned_len; + struct ib_conn *ib_conn = &iser_task->iser_conn->ib_conn; + struct iser_device *device = ib_conn->device; + struct iser_page_vec *page_vec = rsc->page_vec; + struct ib_fmr_pool *fmr_pool = rsc->fmr_pool; + struct ib_pool_fmr *fmr; + int ret, plen; + + plen = iser_sg_to_page_vec(mem, device->ib_device, + page_vec->pages, + &page_vec->offset, + &page_vec->data_size); + page_vec->length = plen; + if (plen * SIZE_4K < page_vec->data_size) { + iser_err("page vec too short to hold this SG\n"); + iser_data_buf_dump(mem, device->ib_device); + iser_dump_page_vec(page_vec); + return -EINVAL; + } + + fmr = ib_fmr_pool_map_phys(fmr_pool, + page_vec->pages, + page_vec->length, + page_vec->pages[0]); + if (IS_ERR(fmr)) { + ret = PTR_ERR(fmr); + iser_err("ib_fmr_pool_map_phys failed: %d\n", ret); + return ret; + } + + reg->sge.lkey = fmr->fmr->lkey; + reg->rkey = fmr->fmr->rkey; + reg->sge.addr = page_vec->pages[0] + page_vec->offset; + reg->sge.length = page_vec->data_size; + reg->mem_h = fmr; + + iser_dbg("fmr reg: 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; +} + +/** + * Unregister (previosuly registered using FMR) memory. + * If memory is non-FMR does nothing. + */ +void iser_unreg_mem_fmr(struct iscsi_iser_task *iser_task, + enum iser_data_dir cmd_dir) +{ + struct iser_mem_reg *reg = &iser_task->rdma_reg[cmd_dir]; + int ret; + + if (!reg->mem_h) + return; + + iser_dbg("PHYSICAL Mem.Unregister mem_h %p\n", reg->mem_h); + + ret = ib_fmr_pool_unmap((struct ib_pool_fmr *)reg->mem_h); + if (ret) + iser_err("ib_fmr_pool_unmap failed %d\n", ret); + + reg->mem_h = NULL; +} + +void iser_unreg_mem_fastreg(struct iscsi_iser_task *iser_task, + enum iser_data_dir cmd_dir) +{ + struct iser_device *device = iser_task->iser_conn->ib_conn.device; + struct iser_mem_reg *reg = &iser_task->rdma_reg[cmd_dir]; + + if (!reg->mem_h) + return; + + device->reg_ops->reg_desc_put(&iser_task->iser_conn->ib_conn, + reg->mem_h); + reg->mem_h = NULL; +} + +static void +iser_set_dif_domain(struct scsi_cmnd *sc, struct ib_sig_attrs *sig_attrs, + struct ib_sig_domain *domain) +{ + domain->sig_type = IB_SIG_TYPE_T10_DIF; + domain->sig.dif.pi_interval = scsi_prot_interval(sc); + domain->sig.dif.ref_tag = scsi_prot_ref_tag(sc); + /* + * At the moment we hard code those, but in the future + * we will take them from sc. + */ + domain->sig.dif.apptag_check_mask = 0xffff; + domain->sig.dif.app_escape = true; + domain->sig.dif.ref_escape = true; + if (sc->prot_flags & SCSI_PROT_REF_INCREMENT) + domain->sig.dif.ref_remap = true; +}; + +static int +iser_set_sig_attrs(struct scsi_cmnd *sc, struct ib_sig_attrs *sig_attrs) +{ + switch (scsi_get_prot_op(sc)) { + case SCSI_PROT_WRITE_INSERT: + case SCSI_PROT_READ_STRIP: + sig_attrs->mem.sig_type = IB_SIG_TYPE_NONE; + iser_set_dif_domain(sc, sig_attrs, &sig_attrs->wire); + sig_attrs->wire.sig.dif.bg_type = IB_T10DIF_CRC; + break; + case SCSI_PROT_READ_INSERT: + case SCSI_PROT_WRITE_STRIP: + sig_attrs->wire.sig_type = IB_SIG_TYPE_NONE; + iser_set_dif_domain(sc, sig_attrs, &sig_attrs->mem); + sig_attrs->mem.sig.dif.bg_type = sc->prot_flags & SCSI_PROT_IP_CHECKSUM ? + IB_T10DIF_CSUM : IB_T10DIF_CRC; + break; + case SCSI_PROT_READ_PASS: + case SCSI_PROT_WRITE_PASS: + iser_set_dif_domain(sc, sig_attrs, &sig_attrs->wire); + sig_attrs->wire.sig.dif.bg_type = IB_T10DIF_CRC; + iser_set_dif_domain(sc, sig_attrs, &sig_attrs->mem); + sig_attrs->mem.sig.dif.bg_type = sc->prot_flags & SCSI_PROT_IP_CHECKSUM ? + IB_T10DIF_CSUM : IB_T10DIF_CRC; + break; + default: + iser_err("Unsupported PI operation %d\n", + scsi_get_prot_op(sc)); + return -EINVAL; + } + + return 0; +} + +static inline void +iser_set_prot_checks(struct scsi_cmnd *sc, u8 *mask) +{ + *mask = 0; + if (sc->prot_flags & SCSI_PROT_REF_CHECK) + *mask |= ISER_CHECK_REFTAG; + if (sc->prot_flags & SCSI_PROT_GUARD_CHECK) + *mask |= ISER_CHECK_GUARD; +} + +static void +iser_inv_rkey(struct ib_send_wr *inv_wr, struct ib_mr *mr) +{ + u32 rkey; + + inv_wr->opcode = IB_WR_LOCAL_INV; + inv_wr->wr_id = ISER_FASTREG_LI_WRID; + inv_wr->ex.invalidate_rkey = mr->rkey; + inv_wr->send_flags = 0; + inv_wr->num_sge = 0; + + rkey = ib_inc_rkey(mr->rkey); + ib_update_fast_reg_key(mr, rkey); +} + +static int +iser_reg_sig_mr(struct iscsi_iser_task *iser_task, + struct iser_pi_context *pi_ctx, + struct iser_mem_reg *data_reg, + struct iser_mem_reg *prot_reg, + struct iser_mem_reg *sig_reg) +{ + struct iser_tx_desc *tx_desc = &iser_task->desc; + struct ib_sig_attrs *sig_attrs = &tx_desc->sig_attrs; + struct ib_sig_handover_wr *wr; + int ret; + + memset(sig_attrs, 0, sizeof(*sig_attrs)); + ret = iser_set_sig_attrs(iser_task->sc, sig_attrs); + if (ret) + goto err; + + iser_set_prot_checks(iser_task->sc, &sig_attrs->check_mask); + + if (!pi_ctx->sig_mr_valid) + iser_inv_rkey(iser_tx_next_wr(tx_desc), pi_ctx->sig_mr); + + wr = sig_handover_wr(iser_tx_next_wr(tx_desc)); + wr->wr.opcode = IB_WR_REG_SIG_MR; + wr->wr.wr_id = ISER_FASTREG_LI_WRID; + wr->wr.sg_list = &data_reg->sge; + wr->wr.num_sge = 1; + wr->wr.send_flags = 0; + wr->sig_attrs = sig_attrs; + wr->sig_mr = pi_ctx->sig_mr; + if (scsi_prot_sg_count(iser_task->sc)) + wr->prot = &prot_reg->sge; + else + wr->prot = NULL; + wr->access_flags = IB_ACCESS_LOCAL_WRITE | + IB_ACCESS_REMOTE_READ | + IB_ACCESS_REMOTE_WRITE; + pi_ctx->sig_mr_valid = 0; + + sig_reg->sge.lkey = pi_ctx->sig_mr->lkey; + sig_reg->rkey = pi_ctx->sig_mr->rkey; + sig_reg->sge.addr = 0; + sig_reg->sge.length = scsi_transfer_length(iser_task->sc); + + iser_dbg("lkey=0x%x rkey=0x%x addr=0x%llx length=%u\n", + sig_reg->sge.lkey, sig_reg->rkey, sig_reg->sge.addr, + sig_reg->sge.length); +err: + return ret; +} + +static int iser_fast_reg_mr(struct iscsi_iser_task *iser_task, + struct iser_data_buf *mem, + struct iser_reg_resources *rsc, + struct iser_mem_reg *reg) +{ + struct iser_tx_desc *tx_desc = &iser_task->desc; + struct ib_mr *mr = rsc->mr; + struct ib_reg_wr *wr; + int n; + + if (!rsc->mr_valid) + iser_inv_rkey(iser_tx_next_wr(tx_desc), mr); + + n = ib_map_mr_sg(mr, mem->sg, mem->size, SIZE_4K); + if (unlikely(n != mem->size)) { + iser_err("failed to map sg (%d/%d)\n", + n, mem->size); + return n < 0 ? n : -EINVAL; + } + + wr = reg_wr(iser_tx_next_wr(tx_desc)); + wr->wr.opcode = IB_WR_REG_MR; + wr->wr.wr_id = ISER_FASTREG_LI_WRID; + wr->wr.send_flags = 0; + wr->wr.num_sge = 0; + wr->mr = mr; + wr->key = mr->rkey; + wr->access = IB_ACCESS_LOCAL_WRITE | + IB_ACCESS_REMOTE_WRITE | + IB_ACCESS_REMOTE_READ; + + rsc->mr_valid = 0; + + reg->sge.lkey = mr->lkey; + reg->rkey = mr->rkey; + reg->sge.addr = mr->iova; + reg->sge.length = mr->length; + + iser_dbg("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 +iser_reg_prot_sg(struct iscsi_iser_task *task, + struct iser_data_buf *mem, + struct iser_fr_desc *desc, + bool use_dma_key, + struct iser_mem_reg *reg) +{ + struct iser_device *device = task->iser_conn->ib_conn.device; + + if (use_dma_key) + return iser_reg_dma(device, mem, reg); + + return device->reg_ops->reg_mem(task, mem, &desc->pi_ctx->rsc, reg); +} + +static int +iser_reg_data_sg(struct iscsi_iser_task *task, + struct iser_data_buf *mem, + struct iser_fr_desc *desc, + bool use_dma_key, + struct iser_mem_reg *reg) +{ + struct iser_device *device = task->iser_conn->ib_conn.device; + + if (use_dma_key) + return iser_reg_dma(device, mem, reg); + + return device->reg_ops->reg_mem(task, mem, &desc->rsc, reg); +} + +int iser_reg_rdma_mem(struct iscsi_iser_task *task, + enum iser_data_dir dir) +{ + struct ib_conn *ib_conn = &task->iser_conn->ib_conn; + struct iser_device *device = ib_conn->device; + struct iser_data_buf *mem = &task->data[dir]; + struct iser_mem_reg *reg = &task->rdma_reg[dir]; + struct iser_mem_reg *data_reg; + struct iser_fr_desc *desc = NULL; + bool use_dma_key; int err; - int i; - struct scatterlist *sg; - regd_buf = &iser_task->rdma_regd[cmd_dir]; + use_dma_key = (mem->dma_nents == 1 && !iser_always_reg && + scsi_get_prot_op(task->sc) == SCSI_PROT_NORMAL); - aligned_len = iser_data_buf_aligned_len(mem, ibdev); - if (aligned_len != mem->dma_nents || - (!ib_conn->fmr_pool && mem->dma_nents > 1)) { - iscsi_conn->fmr_unalign_cnt++; - iser_warn("rdma alignment violation (%d/%d aligned) or FMR not supported\n", - aligned_len, mem->size); - iser_data_buf_dump(mem, ibdev); - - /* unmap the command data before accessing it */ - iser_dma_unmap_task_data(iser_task); - - /* allocate copy buf, if we are writing, copy the */ - /* unaligned scatterlist, dma map the copy */ - if (iser_start_rdma_unaligned_sg(iser_task, cmd_dir) != 0) - return -ENOMEM; - mem = &iser_task->data_copy[cmd_dir]; + if (!use_dma_key) { + desc = device->reg_ops->reg_desc_get(ib_conn); + reg->mem_h = desc; } - /* 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 && err != -EAGAIN) { - iser_data_buf_dump(mem, ibdev); - iser_err("mem->dma_nents = %d (dlength = 0x%x)\n", - mem->dma_nents, - ntoh24(iser_task->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 ; ipage_vec->length ; i++) - iser_err("page_vec[%d] = 0x%llx\n", i, - (unsigned long long) ib_conn->page_vec->pages[i]); + if (scsi_get_prot_op(task->sc) == SCSI_PROT_NORMAL) + data_reg = reg; + else + data_reg = &task->desc.data_reg; + + err = iser_reg_data_sg(task, mem, desc, use_dma_key, data_reg); + if (unlikely(err)) + goto err_reg; + + if (scsi_get_prot_op(task->sc) != SCSI_PROT_NORMAL) { + struct iser_mem_reg *prot_reg = &task->desc.prot_reg; + + if (scsi_prot_sg_count(task->sc)) { + mem = &task->prot[dir]; + err = iser_reg_prot_sg(task, mem, desc, + use_dma_key, prot_reg); + if (unlikely(err)) + goto err_reg; } - if (err) - return err; + + err = iser_reg_sig_mr(task, desc->pi_ctx, data_reg, + prot_reg, reg); + if (unlikely(err)) + goto err_reg; + + desc->pi_ctx->sig_protected = 1; } + return 0; + +err_reg: + if (desc) + device->reg_ops->reg_desc_put(ib_conn, desc); + + return err; +} + +void iser_unreg_rdma_mem(struct iscsi_iser_task *task, + enum iser_data_dir dir) +{ + struct iser_device *device = task->iser_conn->ib_conn.device; + + device->reg_ops->unreg_mem(task, dir); }