This is a note to let you know that I've just added the patch titled
RDMA/hns: Create QP with selected QPN for bank load balance
to the 5.10-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary
The filename of the patch is:
rdma-hns-create-qp-with-selected-qpn-for-bank-load-b.patch
and it can be found in the queue-5.10 subdirectory.
If you, or anyone else, feels it should not be added to the stable tree,
please let <stable@xxxxxxxxxxxxxxx> know about it.
commit 35e78f10a28d9677b66661c45258c791bfe9a10d
Author: Yangyang Li <liyangyang20@xxxxxxxxxx>
Date: Tue Nov 24 20:24:09 2020 +0800
RDMA/hns: Create QP with selected QPN for bank load balance
[ Upstream commit 71586dd2001087e89e344e2c7dcee6b4a53bb6de ]
In order to improve performance by balancing the load between different
banks of cache, the QPC cache is desigend to choose one of 8 banks
according to lower 3 bits of QPN. The hns driver needs to count the number
of QP on each bank and then assigns the QP being created to the bank with
the minimum load first.
Link: https://lore.kernel.org/r/1606220649-1465-1-git-send-email-liweihang@xxxxxxxxxx
Signed-off-by: Yangyang Li <liyangyang20@xxxxxxxxxx>
Signed-off-by: Weihang Li <liweihang@xxxxxxxxxx>
Signed-off-by: Jason Gunthorpe <jgg@xxxxxxxxxx>
Stable-dep-of: 203b70fda634 ("RDMA/hns: Fix return value in hns_roce_map_mr_sg")
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
diff --git a/drivers/infiniband/hw/hns/hns_roce_device.h b/drivers/infiniband/hw/hns/hns_roce_device.h
index e1f5b92596824..5f16173566614 100644
--- a/drivers/infiniband/hw/hns/hns_roce_device.h
+++ b/drivers/infiniband/hw/hns/hns_roce_device.h
@@ -115,6 +115,8 @@
#define HNS_ROCE_IDX_QUE_ENTRY_SZ 4
#define SRQ_DB_REG 0x230
+#define HNS_ROCE_QP_BANK_NUM 8
+
/* The chip implementation of the consumer index is calculated
* according to twice the actual EQ depth
*/
@@ -520,13 +522,22 @@ struct hns_roce_uar_table {
struct hns_roce_bitmap bitmap;
};
+struct hns_roce_bank {
+ struct ida ida;
+ u32 inuse; /* Number of IDs allocated */
+ u32 min; /* Lowest ID to allocate. */
+ u32 max; /* Highest ID to allocate. */
+ u32 next; /* Next ID to allocate. */
+};
+
struct hns_roce_qp_table {
- struct hns_roce_bitmap bitmap;
struct hns_roce_hem_table qp_table;
struct hns_roce_hem_table irrl_table;
struct hns_roce_hem_table trrl_table;
struct hns_roce_hem_table sccc_table;
struct mutex scc_mutex;
+ struct hns_roce_bank bank[HNS_ROCE_QP_BANK_NUM];
+ spinlock_t bank_lock;
};
struct hns_roce_cq_table {
@@ -776,7 +787,7 @@ struct hns_roce_caps {
u32 max_rq_sg;
u32 max_extend_sg;
int num_qps;
- int reserved_qps;
+ u32 reserved_qps;
int num_qpc_timer;
int num_cqc_timer;
int num_srqs;
diff --git a/drivers/infiniband/hw/hns/hns_roce_qp.c b/drivers/infiniband/hw/hns/hns_roce_qp.c
index d1c07f1f8fe98..8323de2c4bf9a 100644
--- a/drivers/infiniband/hw/hns/hns_roce_qp.c
+++ b/drivers/infiniband/hw/hns/hns_roce_qp.c
@@ -154,9 +154,50 @@ static void hns_roce_ib_qp_event(struct hns_roce_qp *hr_qp,
}
}
+static u8 get_least_load_bankid_for_qp(struct hns_roce_bank *bank)
+{
+ u32 least_load = bank[0].inuse;
+ u8 bankid = 0;
+ u32 bankcnt;
+ u8 i;
+
+ for (i = 1; i < HNS_ROCE_QP_BANK_NUM; i++) {
+ bankcnt = bank[i].inuse;
+ if (bankcnt < least_load) {
+ least_load = bankcnt;
+ bankid = i;
+ }
+ }
+
+ return bankid;
+}
+
+static int alloc_qpn_with_bankid(struct hns_roce_bank *bank, u8 bankid,
+ unsigned long *qpn)
+{
+ int id;
+
+ id = ida_alloc_range(&bank->ida, bank->next, bank->max, GFP_KERNEL);
+ if (id < 0) {
+ id = ida_alloc_range(&bank->ida, bank->min, bank->max,
+ GFP_KERNEL);
+ if (id < 0)
+ return id;
+ }
+
+ /* the QPN should keep increasing until the max value is reached. */
+ bank->next = (id + 1) > bank->max ? bank->min : id + 1;
+
+ /* the lower 3 bits is bankid */
+ *qpn = (id << 3) | bankid;
+
+ return 0;
+}
static int alloc_qpn(struct hns_roce_dev *hr_dev, struct hns_roce_qp *hr_qp)
{
+ struct hns_roce_qp_table *qp_table = &hr_dev->qp_table;
unsigned long num = 0;
+ u8 bankid;
int ret;
if (hr_qp->ibqp.qp_type == IB_QPT_GSI) {
@@ -169,13 +210,21 @@ static int alloc_qpn(struct hns_roce_dev *hr_dev, struct hns_roce_qp *hr_qp)
hr_qp->doorbell_qpn = 1;
} else {
- ret = hns_roce_bitmap_alloc_range(&hr_dev->qp_table.bitmap,
- 1, 1, &num);
+ spin_lock(&qp_table->bank_lock);
+ bankid = get_least_load_bankid_for_qp(qp_table->bank);
+
+ ret = alloc_qpn_with_bankid(&qp_table->bank[bankid], bankid,
+ &num);
if (ret) {
- ibdev_err(&hr_dev->ib_dev, "Failed to alloc bitmap\n");
- return -ENOMEM;
+ ibdev_err(&hr_dev->ib_dev,
+ "failed to alloc QPN, ret = %d\n", ret);
+ spin_unlock(&qp_table->bank_lock);
+ return ret;
}
+ qp_table->bank[bankid].inuse++;
+ spin_unlock(&qp_table->bank_lock);
+
hr_qp->doorbell_qpn = (u32)num;
}
@@ -340,9 +389,15 @@ static void free_qpc(struct hns_roce_dev *hr_dev, struct hns_roce_qp *hr_qp)
hns_roce_table_put(hr_dev, &qp_table->irrl_table, hr_qp->qpn);
}
+static inline u8 get_qp_bankid(unsigned long qpn)
+{
+ /* The lower 3 bits of QPN are used to hash to different banks */
+ return (u8)(qpn & GENMASK(2, 0));
+}
+
static void free_qpn(struct hns_roce_dev *hr_dev, struct hns_roce_qp *hr_qp)
{
- struct hns_roce_qp_table *qp_table = &hr_dev->qp_table;
+ u8 bankid;
if (hr_qp->ibqp.qp_type == IB_QPT_GSI)
return;
@@ -350,7 +405,13 @@ static void free_qpn(struct hns_roce_dev *hr_dev, struct hns_roce_qp *hr_qp)
if (hr_qp->qpn < hr_dev->caps.reserved_qps)
return;
- hns_roce_bitmap_free_range(&qp_table->bitmap, hr_qp->qpn, 1, BITMAP_RR);
+ bankid = get_qp_bankid(hr_qp->qpn);
+
+ ida_free(&hr_dev->qp_table.bank[bankid].ida, hr_qp->qpn >> 3);
+
+ spin_lock(&hr_dev->qp_table.bank_lock);
+ hr_dev->qp_table.bank[bankid].inuse--;
+ spin_unlock(&hr_dev->qp_table.bank_lock);
}
static int set_rq_size(struct hns_roce_dev *hr_dev, struct ib_qp_cap *cap,
@@ -1293,22 +1354,24 @@ bool hns_roce_wq_overflow(struct hns_roce_wq *hr_wq, int nreq,
int hns_roce_init_qp_table(struct hns_roce_dev *hr_dev)
{
struct hns_roce_qp_table *qp_table = &hr_dev->qp_table;
- int reserved_from_top = 0;
- int reserved_from_bot;
- int ret;
+ unsigned int reserved_from_bot;
+ unsigned int i;
mutex_init(&qp_table->scc_mutex);
xa_init(&hr_dev->qp_table_xa);
reserved_from_bot = hr_dev->caps.reserved_qps;
- ret = hns_roce_bitmap_init(&qp_table->bitmap, hr_dev->caps.num_qps,
- hr_dev->caps.num_qps - 1, reserved_from_bot,
- reserved_from_top);
- if (ret) {
- dev_err(hr_dev->dev, "qp bitmap init failed!error=%d\n",
- ret);
- return ret;
+ for (i = 0; i < reserved_from_bot; i++) {
+ hr_dev->qp_table.bank[get_qp_bankid(i)].inuse++;
+ hr_dev->qp_table.bank[get_qp_bankid(i)].min++;
+ }
+
+ for (i = 0; i < HNS_ROCE_QP_BANK_NUM; i++) {
+ ida_init(&hr_dev->qp_table.bank[i].ida);
+ hr_dev->qp_table.bank[i].max = hr_dev->caps.num_qps /
+ HNS_ROCE_QP_BANK_NUM - 1;
+ hr_dev->qp_table.bank[i].next = hr_dev->qp_table.bank[i].min;
}
return 0;
@@ -1316,5 +1379,8 @@ int hns_roce_init_qp_table(struct hns_roce_dev *hr_dev)
void hns_roce_cleanup_qp_table(struct hns_roce_dev *hr_dev)
{
- hns_roce_bitmap_cleanup(&hr_dev->qp_table.bitmap);
+ int i;
+
+ for (i = 0; i < HNS_ROCE_QP_BANK_NUM; i++)
+ ida_destroy(&hr_dev->qp_table.bank[i].ida);
}
