RDMA/hns: Bugfix for checking whether the srq is full when post wr

If a user posts WR by wr_list, the head pointer of idx_queue won't be
updated until all wqes are filled, so the judgment of whether head equals
to tail will get a wrong result. Fix above issue and move the head and
tail pointer from the srq structure into the idx_queue structure. After
idx_queue is filled with wqe idx, the head pointer of it will increase.

Fixes: c7bcb13442 ("RDMA/hns: Add SRQ support for hip08 kernel mode")
Link: https://lore.kernel.org/r/1611997090-48820-3-git-send-email-liweihang@huawei.com
Signed-off-by: Wenpeng Liang <liangwenpeng@huawei.com>
Signed-off-by: Weihang Li <liweihang@huawei.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
This commit is contained in:
Wenpeng Liang 2021-01-30 16:58:00 +08:00 коммит произвёл Jason Gunthorpe
Родитель 9dd052474a
Коммит 1620f09b96
3 изменённых файлов: 19 добавлений и 9 удалений

Просмотреть файл

@ -494,6 +494,8 @@ struct hns_roce_idx_que {
struct hns_roce_mtr mtr;
int entry_shift;
unsigned long *bitmap;
u32 head;
u32 tail;
};
struct hns_roce_srq {
@ -513,8 +515,6 @@ struct hns_roce_srq {
u64 *wrid;
struct hns_roce_idx_que idx_que;
spinlock_t lock;
u16 head;
u16 tail;
struct mutex mutex;
void (*event)(struct hns_roce_srq *srq, enum hns_roce_event event);
};

Просмотреть файл

@ -849,11 +849,20 @@ static void hns_roce_free_srq_wqe(struct hns_roce_srq *srq, int wqe_index)
spin_lock(&srq->lock);
bitmap_clear(srq->idx_que.bitmap, wqe_index, 1);
srq->tail++;
srq->idx_que.tail++;
spin_unlock(&srq->lock);
}
int hns_roce_srqwq_overflow(struct hns_roce_srq *srq, int nreq)
{
struct hns_roce_idx_que *idx_que = &srq->idx_que;
unsigned int cur;
cur = idx_que->head - idx_que->tail;
return cur + nreq >= srq->wqe_cnt - 1;
}
static int find_empty_entry(struct hns_roce_idx_que *idx_que,
unsigned long size)
{
@ -889,7 +898,7 @@ static int hns_roce_v2_post_srq_recv(struct ib_srq *ibsrq,
spin_lock_irqsave(&srq->lock, flags);
ind = srq->head & (srq->wqe_cnt - 1);
ind = srq->idx_que.head & (srq->wqe_cnt - 1);
max_sge = srq->max_gs - srq->rsv_sge;
for (nreq = 0; wr; ++nreq, wr = wr->next) {
@ -902,7 +911,7 @@ static int hns_roce_v2_post_srq_recv(struct ib_srq *ibsrq,
break;
}
if (unlikely(srq->head == srq->tail)) {
if (unlikely(hns_roce_srqwq_overflow(srq, nreq))) {
ret = -ENOMEM;
*bad_wr = wr;
break;
@ -938,7 +947,7 @@ static int hns_roce_v2_post_srq_recv(struct ib_srq *ibsrq,
}
if (likely(nreq)) {
srq->head += nreq;
srq->idx_que.head += nreq;
/*
* Make sure that descriptors are written before
@ -950,7 +959,7 @@ static int hns_roce_v2_post_srq_recv(struct ib_srq *ibsrq,
cpu_to_le32(HNS_ROCE_V2_SRQ_DB << V2_DB_BYTE_4_CMD_S |
(srq->srqn & V2_DB_BYTE_4_TAG_M));
srq_db.parameter =
cpu_to_le32(srq->head & V2_DB_PARAMETER_IDX_M);
cpu_to_le32(srq->idx_que.head & V2_DB_PARAMETER_IDX_M);
hns_roce_write64(hr_dev, (__le32 *)&srq_db, srq->db_reg_l);
}

Просмотреть файл

@ -245,6 +245,9 @@ static int alloc_srq_idx(struct hns_roce_dev *hr_dev, struct hns_roce_srq *srq,
}
}
idx_que->head = 0;
idx_que->tail = 0;
return 0;
err_idx_mtr:
hns_roce_mtr_destroy(hr_dev, &idx_que->mtr);
@ -263,8 +266,6 @@ static void free_srq_idx(struct hns_roce_dev *hr_dev, struct hns_roce_srq *srq)
static int alloc_srq_wrid(struct hns_roce_dev *hr_dev, struct hns_roce_srq *srq)
{
srq->head = 0;
srq->tail = srq->wqe_cnt - 1;
srq->wrid = kvmalloc_array(srq->wqe_cnt, sizeof(u64), GFP_KERNEL);
if (!srq->wrid)
return -ENOMEM;