Unverified Commit 85753087 authored by openeuler-ci-bot's avatar openeuler-ci-bot Committed by Gitee
Browse files

!5541 RDMA/hns: Backport bugfixes 

Merge Pull Request from: @stinft 
 
Chengchang Tang (3):
   RDMA/hns: Support flexible wqe buffer page size
   RDMA/hns: Fix missing reset notification by user space driver
   RDMA/hns: Refactor hns_roce_alloc_ucontext()
 Yixing Liu (1):
   RDMA/hns: Kernel notify usr space to stop ring db
https://gitee.com/openeuler/kernel/issues/I98HIN
https://gitee.com/openeuler/kernel/issues/I98HQV
https://gitee.com/openeuler/kernel/issues/I98RAX 
 
Link:https://gitee.com/openeuler/kernel/pulls/5541

 

Reviewed-by: default avatarChengchang Tang <tangchengchang@huawei.com>
Signed-off-by: default avatarZheng Zengkai <zhengzengkai@huawei.com>
parents d72a04ac 2efafc66

drivers/infiniband/hw/hns/hns_roce_device.h

+5 −0
Original line number Diff line number Diff line
@@ -61,6 +61,7 @@ 


#define HNS_ROCE_CEQ				0

#define HNS_ROCE_AEQ				1

#define HNS_ROCE_IS_RESETTING			1



#define HNS_ROCE_CEQE_SIZE 0x4

#define HNS_ROCE_AEQE_SIZE 0x10
@@ -197,6 +198,7 @@ struct hns_roce_uar { 
enum hns_roce_mmap_type {

	HNS_ROCE_MMAP_TYPE_DB = 1,

	HNS_ROCE_MMAP_TYPE_DWQE,

	HNS_ROCE_MMAP_TYPE_RESET = 4,

};



struct hns_user_mmap_entry {
@@ -211,6 +213,7 @@ struct hns_roce_ucontext { 
	struct list_head	page_list;

	struct mutex		page_mutex;

	struct hns_user_mmap_entry *db_mmap_entry;

	struct hns_user_mmap_entry *reset_mmap_entry;

	u32			config;

};
@@ -1041,6 +1044,8 @@ struct hns_roce_dev { 
	int			loop_idc;

	u32			sdb_offset;

	u32			odb_offset;

	struct page		*reset_page; /* store reset state */

	void			*reset_kaddr; /* addr of reset page */

	const struct hns_roce_hw *hw;

	void			*priv;

	struct workqueue_struct *irq_workq;

drivers/infiniband/hw/hns/hns_roce_hw_v2.c

+59 −2
Original line number Diff line number Diff line
@@ -2607,6 +2607,36 @@ static void free_dip_list(struct hns_roce_dev *hr_dev) 
	spin_unlock_irqrestore(&hr_dev->dip_list_lock, flags);

}



static int hns_roce_v2_get_reset_page(struct hns_roce_dev *hr_dev)

{

	struct hns_roce_v2_reset_state *state;



	hr_dev->reset_page = alloc_page(GFP_KERNEL | __GFP_ZERO);

	if (!hr_dev->reset_page)

		return -ENOMEM;



	hr_dev->reset_kaddr = vmap(&hr_dev->reset_page, 1, VM_MAP, PAGE_KERNEL);

	if (!hr_dev->reset_kaddr)

		goto err_with_vmap;



	state = hr_dev->reset_kaddr;

	state->hw_ready = ~state->hw_ready;



	return 0;



err_with_vmap:

	put_page(hr_dev->reset_page);

	return -ENOMEM;

}



static void hns_roce_v2_put_reset_page(struct hns_roce_dev *hr_dev)

{

	vunmap(hr_dev->reset_kaddr);

	hr_dev->reset_kaddr = NULL;

	put_page(hr_dev->reset_page);

	hr_dev->reset_page = NULL;

}



static struct ib_pd *free_mr_init_pd(struct hns_roce_dev *hr_dev)

{

	struct hns_roce_v2_priv *priv = hr_dev->priv;
@@ -2971,14 +3001,21 @@ static int hns_roce_v2_init(struct hns_roce_dev *hr_dev) 
{

	int ret;



	ret = hns_roce_v2_get_reset_page(hr_dev);

	if (ret) {

		dev_err(hr_dev->dev,

			"reset state init failed, ret = %d.\n", ret);

		return ret;

	}



	/* The hns ROCEE requires the extdb info to be cleared before using */

	ret = hns_roce_clear_extdb_list_info(hr_dev);

	if (ret)

		return ret;

		goto err_clear_extdb_failed;



	ret = get_hem_table(hr_dev);

	if (ret)

		return ret;

		goto err_clear_extdb_failed;



	if (hr_dev->is_vf)

		return 0;
@@ -2993,6 +3030,8 @@ static int hns_roce_v2_init(struct hns_roce_dev *hr_dev) 


err_llm_init_failed:

	put_hem_table(hr_dev);

err_clear_extdb_failed:

	hns_roce_v2_put_reset_page(hr_dev);



	return ret;

}
@@ -3004,6 +3043,8 @@ static void hns_roce_v2_exit(struct hns_roce_dev *hr_dev) 
	if (!hr_dev->is_vf)

		hns_roce_free_link_table(hr_dev);



	hns_roce_v2_put_reset_page(hr_dev);



	if (hr_dev->pci_dev->revision == PCI_REVISION_ID_HIP09)

		free_dip_list(hr_dev);

}
@@ -7165,6 +7206,19 @@ int hns_roce_bond_uninit_client(struct hns_roce_bond_group *bond_grp, 


	return 0;

}



static void hns_roce_v2_reset_notify_user(struct hns_roce_dev *hr_dev)

{

	struct hns_roce_v2_reset_state *state;



	state = (struct hns_roce_v2_reset_state *)hr_dev->reset_kaddr;



	state->reset_state = HNS_ROCE_IS_RESETTING;

	state->hw_ready = 0;

	/* Ensure reset state was flushed in memory */

	wmb();

}



static int hns_roce_hw_v2_reset_notify_down(struct hnae3_handle *handle)

{

	struct hns_roce_dev *hr_dev;
@@ -7183,6 +7237,9 @@ static int hns_roce_hw_v2_reset_notify_down(struct hnae3_handle *handle) 


	hr_dev->active = false;

	hr_dev->dis_db = true;



	hns_roce_v2_reset_notify_user(hr_dev);



	hr_dev->state = HNS_ROCE_DEVICE_STATE_RST_DOWN;



	return 0;

drivers/infiniband/hw/hns/hns_roce_hw_v2.h

+5 −0
Original line number Diff line number Diff line
@@ -1313,6 +1313,11 @@ struct hns_roce_link_table { 
#define HNS_ROCE_EXT_LLM_ENTRY(addr, id) (((id) << (64 - 12)) | ((addr) >> 12))

#define HNS_ROCE_EXT_LLM_MIN_PAGES(que_num) ((que_num) * 4 + 2)



struct hns_roce_v2_reset_state {

	u32 reset_state; /* stored to use in user space */

	u32 hw_ready;

};



struct hns_roce_v2_free_mr {

	struct hns_roce_qp *rsv_qp[HNS_ROCE_FREE_MR_USED_QP_NUM];

	struct hns_roce_cq *rsv_cq;

drivers/infiniband/hw/hns/hns_roce_main.c

+77 −20
Original line number Diff line number Diff line
@@ -392,6 +392,7 @@ hns_roce_user_mmap_entry_insert(struct ib_ucontext *ucontext, u64 address, 
				ucontext, &entry->rdma_entry, length, 0);

		break;

	case HNS_ROCE_MMAP_TYPE_DWQE:

	case HNS_ROCE_MMAP_TYPE_RESET:

		ret = rdma_user_mmap_entry_insert_range(

				ucontext, &entry->rdma_entry, length, 1,

				U32_MAX);
@@ -430,49 +431,83 @@ static int hns_roce_alloc_uar_entry(struct ib_ucontext *uctx) 
	return 0;

}



static int hns_roce_alloc_ucontext(struct ib_ucontext *uctx,

				   struct ib_udata *udata)

static void hns_roce_dealloc_reset_entry(struct hns_roce_ucontext *context)

{

	if (context->reset_mmap_entry)

		rdma_user_mmap_entry_remove(&context->reset_mmap_entry->rdma_entry);

}



static int hns_roce_alloc_reset_entry(struct ib_ucontext *uctx)

{

	struct hns_roce_ucontext *context = to_hr_ucontext(uctx);

	struct hns_roce_dev *hr_dev = to_hr_dev(uctx->device);

	struct hns_roce_ib_alloc_ucontext_resp resp = {};

	struct hns_roce_ib_alloc_ucontext ucmd = {};

	int ret = -EAGAIN;



	if (!hr_dev->active)

		goto error_out;

	context->reset_mmap_entry = hns_roce_user_mmap_entry_insert(uctx,

		(u64)hr_dev->reset_kaddr, PAGE_SIZE, HNS_ROCE_MMAP_TYPE_RESET);



	resp.qp_tab_size = hr_dev->caps.num_qps;

	resp.srq_tab_size = hr_dev->caps.num_srqs;

	if (!context->reset_mmap_entry)

		return -ENOMEM;



	ret = ib_copy_from_udata(&ucmd, udata,

				 min(udata->inlen, sizeof(ucmd)));

	if (ret)

		goto error_out;

	return 0;

}



static void hns_roce_get_uctx_config(struct hns_roce_dev *hr_dev,

				struct hns_roce_ucontext *context,

				struct hns_roce_ib_alloc_ucontext *ucmd,

				struct hns_roce_ib_alloc_ucontext_resp *resp)

{

	if (hr_dev->pci_dev->revision >= PCI_REVISION_ID_HIP09)

		context->config = ucmd.config & HNS_ROCE_EXSGE_FLAGS;

		context->config = ucmd->config & HNS_ROCE_EXSGE_FLAGS;



	if (context->config & HNS_ROCE_EXSGE_FLAGS) {

		resp.config |= HNS_ROCE_RSP_EXSGE_FLAGS;

		resp.max_inline_data = hr_dev->caps.max_sq_inline;

		resp->config |= HNS_ROCE_RSP_EXSGE_FLAGS;

		resp->max_inline_data = hr_dev->caps.max_sq_inline;

	}



	if (hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_RQ_INLINE) {

		context->config |= ucmd.config & HNS_ROCE_RQ_INLINE_FLAGS;

		context->config |= ucmd->config & HNS_ROCE_RQ_INLINE_FLAGS;

		if (context->config & HNS_ROCE_RQ_INLINE_FLAGS)

			resp.config |= HNS_ROCE_RSP_RQ_INLINE_FLAGS;

			resp->config |= HNS_ROCE_RSP_RQ_INLINE_FLAGS;

	}



	if (hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_CQE_INLINE) {

		context->config |= ucmd.config & HNS_ROCE_CQE_INLINE_FLAGS;

		context->config |= ucmd->config & HNS_ROCE_CQE_INLINE_FLAGS;

		if (context->config & HNS_ROCE_CQE_INLINE_FLAGS)

			resp.config |= HNS_ROCE_RSP_CQE_INLINE_FLAGS;

			resp->config |= HNS_ROCE_RSP_CQE_INLINE_FLAGS;

	}



	if (ucmd->config & HNS_ROCE_UCTX_DYN_QP_PGSZ) {

		context->config |= HNS_ROCE_UCTX_DYN_QP_PGSZ;

		resp->config |=  HNS_ROCE_UCTX_RSP_DYN_QP_PGSZ;

	}

}



static int hns_roce_alloc_ucontext(struct ib_ucontext *uctx,

				   struct ib_udata *udata)

{

	struct hns_roce_ucontext *context = to_hr_ucontext(uctx);

	struct hns_roce_dev *hr_dev = to_hr_dev(uctx->device);

	struct hns_roce_ib_alloc_ucontext_resp resp = {};

	struct hns_roce_ib_alloc_ucontext ucmd = {};

	struct rdma_user_mmap_entry *rdma_entry;

	int ret = -EAGAIN;



	if (!hr_dev->active)

		goto error_out;



	resp.qp_tab_size = hr_dev->caps.num_qps;

	resp.srq_tab_size = hr_dev->caps.num_srqs;



	ret = ib_copy_from_udata(&ucmd, udata,

				 min(udata->inlen, sizeof(ucmd)));

	if (ret)

		goto error_out;



	if (hr_dev->pci_dev->revision >= PCI_REVISION_ID_HIP09)

		resp.congest_type = hr_dev->caps.cong_cap;



	hns_roce_get_uctx_config(hr_dev, context, &ucmd, &resp);



	ret = hns_roce_uar_alloc(hr_dev, &context->uar);

	if (ret)

		goto error_out;
@@ -487,6 +522,15 @@ static int hns_roce_alloc_ucontext(struct ib_ucontext *uctx, 
		mutex_init(&context->page_mutex);

	}



	ret = hns_roce_alloc_reset_entry(uctx);

	if (ret)

		goto error_fail_reset_entry;



	if (context->reset_mmap_entry) {

		rdma_entry = &context->reset_mmap_entry->rdma_entry;

		resp.reset_mmap_key = rdma_user_mmap_get_offset(rdma_entry);

	}



	resp.cqe_size = hr_dev->caps.cqe_sz;



	ret = ib_copy_to_udata(udata, &resp,
@@ -497,6 +541,9 @@ static int hns_roce_alloc_ucontext(struct ib_ucontext *uctx, 
	return 0;



error_fail_copy_to_udata:

	hns_roce_dealloc_reset_entry(context);



error_fail_reset_entry:

	hns_roce_dealloc_uar_entry(context);



error_fail_uar_entry:
@@ -514,6 +561,7 @@ static void hns_roce_dealloc_ucontext(struct ib_ucontext *ibcontext) 
	struct hns_roce_dev *hr_dev = to_hr_dev(ibcontext->device);



	hns_roce_dealloc_uar_entry(context);

	hns_roce_dealloc_reset_entry(context);



	ida_free(&hr_dev->uar_ida.ida, (int)context->uar.logic_idx);

}
@@ -541,6 +589,15 @@ static int hns_roce_mmap(struct ib_ucontext *uctx, struct vm_area_struct *vma) 
	case HNS_ROCE_MMAP_TYPE_DWQE:

		prot = pgprot_device(vma->vm_page_prot);

		break;

	case HNS_ROCE_MMAP_TYPE_RESET:

		if (vma->vm_flags & (VM_WRITE | VM_EXEC)) {

			ret = -EINVAL;

			goto out;

		}



		prot = vma->vm_page_prot;

		pfn = page_to_pfn(hr_dev->reset_page);

		break;

	default:

		ret = -EINVAL;

		goto out;

drivers/infiniband/hw/hns/hns_roce_qp.c

+20 −12
Original line number Diff line number Diff line
@@ -641,18 +641,21 @@ static int set_user_sq_size(struct hns_roce_dev *hr_dev, 
}



static int set_wqe_buf_attr(struct hns_roce_dev *hr_dev,

			    struct hns_roce_qp *hr_qp,

			    struct hns_roce_qp *hr_qp, u8 page_shift,

			    struct hns_roce_buf_attr *buf_attr)

{

	unsigned int page_size = BIT(page_shift);

	int buf_size;

	int idx = 0;



	hr_qp->buff_size = 0;



	if (page_shift > PAGE_SHIFT || page_shift < HNS_HW_PAGE_SHIFT)

		return -EOPNOTSUPP;



	/* SQ WQE */

	hr_qp->sq.offset = 0;

	buf_size = to_hr_hem_entries_size(hr_qp->sq.wqe_cnt,

					  hr_qp->sq.wqe_shift);

	buf_size = ALIGN(hr_qp->sq.wqe_cnt << hr_qp->sq.wqe_shift, page_size);

	if (buf_size > 0 && idx < ARRAY_SIZE(buf_attr->region)) {

		buf_attr->region[idx].size = buf_size;

		buf_attr->region[idx].hopnum = hr_dev->caps.wqe_sq_hop_num;
@@ -662,8 +665,7 @@ static int set_wqe_buf_attr(struct hns_roce_dev *hr_dev, 


	/* extend SGE WQE in SQ */

	hr_qp->sge.offset = hr_qp->buff_size;

	buf_size = to_hr_hem_entries_size(hr_qp->sge.sge_cnt,

					  hr_qp->sge.sge_shift);

	buf_size = ALIGN(hr_qp->sge.sge_cnt << hr_qp->sge.sge_shift, page_size);

	if (buf_size > 0 && idx < ARRAY_SIZE(buf_attr->region)) {

		buf_attr->region[idx].size = buf_size;

		buf_attr->region[idx].hopnum = hr_dev->caps.wqe_sge_hop_num;
@@ -673,8 +675,7 @@ static int set_wqe_buf_attr(struct hns_roce_dev *hr_dev, 


	/* RQ WQE */

	hr_qp->rq.offset = hr_qp->buff_size;

	buf_size = to_hr_hem_entries_size(hr_qp->rq.wqe_cnt,

					  hr_qp->rq.wqe_shift);

	buf_size = ALIGN(hr_qp->rq.wqe_cnt << hr_qp->rq.wqe_shift, page_size);

	if (buf_size > 0 && idx < ARRAY_SIZE(buf_attr->region)) {

		buf_attr->region[idx].size = buf_size;

		buf_attr->region[idx].hopnum = hr_dev->caps.wqe_rq_hop_num;
@@ -685,8 +686,8 @@ static int set_wqe_buf_attr(struct hns_roce_dev *hr_dev, 
	if (hr_qp->buff_size < 1)

		return -EINVAL;



	buf_attr->page_shift = HNS_HW_PAGE_SHIFT + hr_dev->caps.mtt_buf_pg_sz;

	buf_attr->region_count = idx;

	buf_attr->page_shift = page_shift;



	return 0;

}
@@ -742,20 +743,27 @@ static int hns_roce_qp_has_rq(struct ib_qp_init_attr *attr) 


static int alloc_qp_buf(struct hns_roce_dev *hr_dev, struct hns_roce_qp *hr_qp,

			struct ib_qp_init_attr *init_attr,

			struct ib_udata *udata, unsigned long addr)

			struct ib_udata *udata,

			struct hns_roce_ib_create_qp *ucmd)

{

	struct hns_roce_ucontext *uctx = rdma_udata_to_drv_context(udata,

					 struct hns_roce_ucontext, ibucontext);

	struct ib_device *ibdev = &hr_dev->ib_dev;

	struct hns_roce_buf_attr buf_attr = {};

	u8 page_shift = HNS_HW_PAGE_SHIFT;

	int ret;



	ret = set_wqe_buf_attr(hr_dev, hr_qp, &buf_attr);

	if (uctx && (uctx->config & HNS_ROCE_UCTX_DYN_QP_PGSZ))

		page_shift = ucmd->pageshift;



	ret = set_wqe_buf_attr(hr_dev, hr_qp, page_shift, &buf_attr);

	if (ret) {

		ibdev_err(ibdev, "failed to split WQE buf, ret = %d.\n", ret);

		goto err_inline;

	}

	ret = hns_roce_mtr_create(hr_dev, &hr_qp->mtr, &buf_attr,

				  PAGE_SHIFT + hr_dev->caps.mtt_ba_pg_sz,

				  udata, addr);

				  udata, ucmd->buf_addr);

	if (ret) {

		ibdev_err(ibdev, "failed to create WQE mtr, ret = %d.\n", ret);

		goto err_inline;
@@ -1151,7 +1159,7 @@ static int hns_roce_create_qp_common(struct hns_roce_dev *hr_dev, 
		}

	}



	ret = alloc_qp_buf(hr_dev, hr_qp, init_attr, udata, ucmd.buf_addr);

	ret = alloc_qp_buf(hr_dev, hr_qp, init_attr, udata, &ucmd);

	if (ret) {

		ibdev_err(ibdev, "failed to alloc QP buffer, ret = %d.\n", ret);

		goto err_buf;