[PATCH V2 net-next 00/12] code optimizations & bugfixes for HNS3 driver

[PATCH V2 net-next 00/12] code optimizations & bugfixes for HNS3 driver

[Huazhong Tan]

This patch-set includes code optimizations and bugfixes for the HNS3
ethernet controller driver.

[patch 1/12 - 3/12] fixes some bugs about the IO path

[patch 4/12 - 6/12] includes some optimization and bugfixes
about mailbox message handling

[patch 7/12 - 12/12] adds misc code optimizations and bugfixes.

Change log:
V1->V2: adds modification on [patch 8/12]

Huazhong Tan (7):
  net: hns3: stop sending keep alive msg when VF command queue needs
    reinit
  net: hns3: use atomic_t replace u32 for arq's count
  net: hns3: use a reserved byte to identify need_resp flag
  net: hns3: not reset TQP in the DOWN while VF resetting
  net: hns3: stop schedule reset service while unloading driver
  net: hns3: fix pause configure fail problem
  net: hns3: prevent double free in hns3_put_ring_config()

Weihang Li (1):
  net: hns3: remove reset after command send failed

Yunsheng Lin (3):
  net: hns3: fix data race between ring->next_to_clean
  net: hns3: fix for TX clean num when cleaning TX BD
  net: hns3: handle the BD info on the last BD of the packet

liuzhongzhu (1):
  net: hns3: extend the loopback state acquisition time

 drivers/net/ethernet/hisilicon/hns3/hclge_mbx.h    |  7 ++-
 drivers/net/ethernet/hisilicon/hns3/hns3_enet.c    | 65 +++++++++++++---------
 drivers/net/ethernet/hisilicon/hns3/hns3_enet.h    |  7 ++-
 .../net/ethernet/hisilicon/hns3/hns3pf/hclge_err.c | 10 ----
 .../ethernet/hisilicon/hns3/hns3pf/hclge_main.c    |  8 ++-
 .../net/ethernet/hisilicon/hns3/hns3pf/hclge_mbx.c |  7 +--
 .../net/ethernet/hisilicon/hns3/hns3pf/hclge_tm.c  |  5 +-
 .../ethernet/hisilicon/hns3/hns3vf/hclgevf_cmd.c   |  2 +-
 .../ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c  | 12 ++--
 .../ethernet/hisilicon/hns3/hns3vf/hclgevf_main.h  |  1 +
 .../ethernet/hisilicon/hns3/hns3vf/hclgevf_mbx.c   |  9 ++-
 11 files changed, 76 insertions(+), 57 deletions(-)

-- 
2.7.4

[PATCH V2 net-next 01/12] net: hns3: fix data race between ring->next_to_clean

[Huazhong Tan]

From: Yunsheng Lin <linyunsheng@huawei.com>

hns3_clean_tx_ring calls hns3_nic_reclaim_one_desc to clean
buffers and set ring->next_to_clean, then hns3_nic_net_xmit
reuses the cleaned buffers. But there are no memory barriers
when buffers gets recycled, so the recycled buffers can be
corrupted.

This patch uses smp_store_release to update ring->next_to_clean
and smp_load_acquire to read ring->next_to_clean to properly
hand off buffers from hns3_clean_tx_ring to hns3_nic_net_xmit.

Fixes: 76ad4f0ee747 ("net: hns3: Add support of HNS3 Ethernet Driver for hip08 SoC")
Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
---
 drivers/net/ethernet/hisilicon/hns3/hns3_enet.c | 14 +++++++++++---
 drivers/net/ethernet/hisilicon/hns3/hns3_enet.h |  7 +++++--
 2 files changed, 16 insertions(+), 5 deletions(-)

diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
index 176d4b9..9fdc87e 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
@@ -2214,14 +2214,22 @@ static void hns3_reuse_buffer(struct hns3_enet_ring *ring, int i)
 static void hns3_nic_reclaim_one_desc(struct hns3_enet_ring *ring, int *bytes,
 				      int *pkts)
 {
-	struct hns3_desc_cb *desc_cb = &ring->desc_cb[ring->next_to_clean];
+	int ntc = ring->next_to_clean;
+	struct hns3_desc_cb *desc_cb;
 
+	desc_cb = &ring->desc_cb[ntc];
 	(*pkts) += (desc_cb->type == DESC_TYPE_SKB);
 	(*bytes) += desc_cb->length;
 	/* desc_cb will be cleaned, after hnae3_free_buffer_detach*/
-	hns3_free_buffer_detach(ring, ring->next_to_clean);
+	hns3_free_buffer_detach(ring, ntc);
 
-	ring_ptr_move_fw(ring, next_to_clean);
+	if (++ntc == ring->desc_num)
+		ntc = 0;
+
+	/* This smp_store_release() pairs with smp_load_acquire() in
+	 * ring_space called by hns3_nic_net_xmit.
+	 */
+	smp_store_release(&ring->next_to_clean, ntc);
 }
 
 static int is_valid_clean_head(struct hns3_enet_ring *ring, int h)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.h b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.h
index cec56a5..2b4f5ea 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.h
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.h
@@ -581,8 +581,11 @@ union l4_hdr_info {
 
 static inline int ring_space(struct hns3_enet_ring *ring)
 {
-	int begin = ring->next_to_clean;
-	int end = ring->next_to_use;
+	/* This smp_load_acquire() pairs with smp_store_release() in
+	 * hns3_nic_reclaim_one_desc called by hns3_clean_tx_ring.
+	 */
+	int begin = smp_load_acquire(&ring->next_to_clean);
+	int end = READ_ONCE(ring->next_to_use);
 
 	return ((end >= begin) ? (ring->desc_num - end + begin) :
 			(begin - end)) - 1;
-- 
2.7.4

[PATCH V2 net-next 02/12] net: hns3: fix for TX clean num when cleaning TX BD

[Huazhong Tan]

From: Yunsheng Lin <linyunsheng@huawei.com>

hns3_desc_unused() returns how many BD have been cleaned, but new
buffer has not been attached to them. The register of
HNS3_RING_RX_RING_FBDNUM_REG returns how many BD need allocating new
buffer to or need to cleaned. So the remaining BD need to be clean
is HNS3_RING_RX_RING_FBDNUM_REG - hns3_desc_unused().

Also, new buffer can not attach to the pending BD when the last BD is
not handled, because memcpy has not been done on the first pending BD.

This patch fixes by subtracting the pending BD num from unused_count
after 'HNS3_RING_RX_RING_FBDNUM_REG - unused_count' is used to calculate
the BD bum need to be clean.

Fixes: e55970950556 ("net: hns3: Add handling of GRO Pkts not fully RX'ed in NAPI poll")
Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
---
 drivers/net/ethernet/hisilicon/hns3/hns3_enet.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
index 9fdc87e..6695b94 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
@@ -2874,7 +2874,7 @@ int hns3_clean_rx_ring(
 {
 #define RCB_NOF_ALLOC_RX_BUFF_ONCE 16
 	int recv_pkts, recv_bds, clean_count, err;
-	int unused_count = hns3_desc_unused(ring) - ring->pending_buf;
+	int unused_count = hns3_desc_unused(ring);
 	struct sk_buff *skb = ring->skb;
 	int num;
 
@@ -2883,6 +2883,7 @@ int hns3_clean_rx_ring(
 
 	recv_pkts = 0, recv_bds = 0, clean_count = 0;
 	num -= unused_count;
+	unused_count -= ring->pending_buf;
 
 	while (recv_pkts < budget && recv_bds < num) {
 		/* Reuse or realloc buffers */
-- 
2.7.4

[PATCH V2 net-next 03/12] net: hns3: handle the BD info on the last BD of the packet

[Huazhong Tan]

From: Yunsheng Lin <linyunsheng@huawei.com>

The bdinfo handled in hns3_handle_bdinfo is only valid on the
last BD of the current packet, currently the bd info may be handled
based on the first BD if the packet has more than two BDs, which
may cause rx error.

This patch fixes it by using the last BD of the current packet in
hns3_handle_bdinfo.

Also, hns3_set_rx_skb_rss_type has used RSS hash value from the last
BD of the current packet, so remove the same last BD calculation in
hns3_set_rx_skb_rss_type and call it from hns3_handle_bdinfo.

Fixes: e55970950556 ("net: hns3: Add handling of GRO Pkts not fully RX'ed in NAPI poll")

Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
---
 drivers/net/ethernet/hisilicon/hns3/hns3_enet.c | 38 +++++++++++++------------
 1 file changed, 20 insertions(+), 18 deletions(-)

diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
index 6695b94..7dc281c 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
@@ -2697,36 +2697,37 @@ static int hns3_set_gro_and_checksum(struct hns3_enet_ring *ring,
 }
 
 static void hns3_set_rx_skb_rss_type(struct hns3_enet_ring *ring,
-				     struct sk_buff *skb)
+				     struct sk_buff *skb, u32 rss_hash)
 {
 	struct hnae3_handle *handle = ring->tqp->handle;
 	enum pkt_hash_types rss_type;
-	struct hns3_desc *desc;
-	int last_bd;
-
-	/* When driver handle the rss type, ring->next_to_clean indicates the
-	 * first descriptor of next packet, need -1 here.
-	 */
-	last_bd = (ring->next_to_clean - 1 + ring->desc_num) % ring->desc_num;
-	desc = &ring->desc[last_bd];
 
-	if (le32_to_cpu(desc->rx.rss_hash))
+	if (rss_hash)
 		rss_type = handle->kinfo.rss_type;
 	else
 		rss_type = PKT_HASH_TYPE_NONE;
 
-	skb_set_hash(skb, le32_to_cpu(desc->rx.rss_hash), rss_type);
+	skb_set_hash(skb, rss_hash, rss_type);
 }
 
-static int hns3_handle_bdinfo(struct hns3_enet_ring *ring, struct sk_buff *skb,
-			      struct hns3_desc *desc)
+static int hns3_handle_bdinfo(struct hns3_enet_ring *ring, struct sk_buff *skb)
 {
 	struct net_device *netdev = ring->tqp->handle->kinfo.netdev;
-	u32 bd_base_info = le32_to_cpu(desc->rx.bd_base_info);
-	u32 l234info = le32_to_cpu(desc->rx.l234_info);
 	enum hns3_pkt_l2t_type l2_frame_type;
+	u32 bd_base_info, l234info;
+	struct hns3_desc *desc;
 	unsigned int len;
-	int ret;
+	int pre_ntc, ret;
+
+	/* bdinfo handled below is only valid on the last BD of the
+	 * current packet, and ring->next_to_clean indicates the first
+	 * descriptor of next packet, so need - 1 below.
+	 */
+	pre_ntc = ring->next_to_clean ? (ring->next_to_clean - 1) :
+					(ring->desc_num - 1);
+	desc = &ring->desc[pre_ntc];
+	bd_base_info = le32_to_cpu(desc->rx.bd_base_info);
+	l234info = le32_to_cpu(desc->rx.l234_info);
 
 	/* Based on hw strategy, the tag offloaded will be stored at
 	 * ot_vlan_tag in two layer tag case, and stored at vlan_tag
@@ -2787,6 +2788,8 @@ static int hns3_handle_bdinfo(struct hns3_enet_ring *ring, struct sk_buff *skb,
 	u64_stats_update_end(&ring->syncp);
 
 	ring->tqp_vector->rx_group.total_bytes += len;
+
+	hns3_set_rx_skb_rss_type(ring, skb, le32_to_cpu(desc->rx.rss_hash));
 	return 0;
 }
 
@@ -2856,14 +2859,13 @@ static int hns3_handle_rx_bd(struct hns3_enet_ring *ring,
 		       ALIGN(ring->pull_len, sizeof(long)));
 	}
 
-	ret = hns3_handle_bdinfo(ring, skb, desc);
+	ret = hns3_handle_bdinfo(ring, skb);
 	if (unlikely(ret)) {
 		dev_kfree_skb_any(skb);
 		return ret;
 	}
 
 	*out_skb = skb;
-	hns3_set_rx_skb_rss_type(ring, skb);
 
 	return 0;
 }
-- 
2.7.4

[PATCH V2 net-next 04/12] net: hns3: stop sending keep alive msg when VF command queue needs reinit

[Huazhong Tan]

HCLGEVF_STATE_CMD_DISABLE is more suitable than
HCLGEVF_STATE_RST_HANDLING to stop sending keep alive msg,
since HCLGEVF_STATE_RST_HANDLING only be set when the reset
task is running.

Fixes: c59a85c07e77 ("net: hns3: stop sending keep alive msg to PF when VF is resetting")
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
---
 drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c
index f9d98f8..7672cab 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c
@@ -1758,7 +1758,7 @@ static void hclgevf_keep_alive_task(struct work_struct *work)
 
 	hdev = container_of(work, struct hclgevf_dev, keep_alive_task);
 
-	if (test_bit(HCLGEVF_STATE_RST_HANDLING, &hdev->state))
+	if (test_bit(HCLGEVF_STATE_CMD_DISABLE, &hdev->state))
 		return;
 
 	ret = hclgevf_send_mbx_msg(hdev, HCLGE_MBX_KEEP_ALIVE, 0, NULL,
-- 
2.7.4

[PATCH V2 net-next 05/12] net: hns3: use atomic_t replace u32 for arq's count

[Huazhong Tan]

Since irq handler and mailbox task will both update arq's count,
so arq's count should use atomic_t instead of u32, otherwise
its value may go wrong finally.

Fixes: 07a0556a3a73 ("net: hns3: Changes to support ARQ(Asynchronous Receive Queue)")
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
---
 drivers/net/ethernet/hisilicon/hns3/hclge_mbx.h          | 2 +-
 drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_cmd.c | 2 +-
 drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_mbx.c | 7 ++++---
 3 files changed, 6 insertions(+), 5 deletions(-)

diff --git a/drivers/net/ethernet/hisilicon/hns3/hclge_mbx.h b/drivers/net/ethernet/hisilicon/hns3/hclge_mbx.h
index 360463a..8b6191f 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hclge_mbx.h
+++ b/drivers/net/ethernet/hisilicon/hns3/hclge_mbx.h
@@ -111,7 +111,7 @@ struct hclgevf_mbx_arq_ring {
 	struct hclgevf_dev *hdev;
 	u32 head;
 	u32 tail;
-	u32 count;
+	atomic_t count;
 	u16 msg_q[HCLGE_MBX_MAX_ARQ_MSG_NUM][HCLGE_MBX_MAX_ARQ_MSG_SIZE];
 };
 
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_cmd.c b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_cmd.c
index 1b428d4..71f356f 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_cmd.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_cmd.c
@@ -340,7 +340,7 @@ int hclgevf_cmd_init(struct hclgevf_dev *hdev)
 	hdev->arq.hdev = hdev;
 	hdev->arq.head = 0;
 	hdev->arq.tail = 0;
-	hdev->arq.count = 0;
+	atomic_set(&hdev->arq.count, 0);
 	hdev->hw.cmq.csq.next_to_clean = 0;
 	hdev->hw.cmq.csq.next_to_use = 0;
 	hdev->hw.cmq.crq.next_to_clean = 0;
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_mbx.c b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_mbx.c
index eb56287..3c22639 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_mbx.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_mbx.c
@@ -212,7 +212,8 @@ void hclgevf_mbx_handler(struct hclgevf_dev *hdev)
 			/* we will drop the async msg if we find ARQ as full
 			 * and continue with next message
 			 */
-			if (hdev->arq.count >= HCLGE_MBX_MAX_ARQ_MSG_NUM) {
+			if (atomic_read(&hdev->arq.count) >=
+			    HCLGE_MBX_MAX_ARQ_MSG_NUM) {
 				dev_warn(&hdev->pdev->dev,
 					 "Async Q full, dropping msg(%d)\n",
 					 req->msg[1]);
@@ -224,7 +225,7 @@ void hclgevf_mbx_handler(struct hclgevf_dev *hdev)
 			memcpy(&msg_q[0], req->msg,
 			       HCLGE_MBX_MAX_ARQ_MSG_SIZE * sizeof(u16));
 			hclge_mbx_tail_ptr_move_arq(hdev->arq);
-			hdev->arq.count++;
+			atomic_inc(&hdev->arq.count);
 
 			hclgevf_mbx_task_schedule(hdev);
 
@@ -317,7 +318,7 @@ void hclgevf_mbx_async_handler(struct hclgevf_dev *hdev)
 		}
 
 		hclge_mbx_head_ptr_move_arq(hdev->arq);
-		hdev->arq.count--;
+		atomic_dec(&hdev->arq.count);
 		msg_q = hdev->arq.msg_q[hdev->arq.head];
 	}
 }
-- 
2.7.4

[PATCH V2 net-next 06/12] net: hns3: use a reserved byte to identify need_resp flag

[Huazhong Tan]

This patch uses a reserved byte in the hclge_mbx_vf_to_pf_cmd
to save the need_resp flag, so when PF received the mailbox,
it can use it to decise whether send a response to VF.

For hclge_set_vf_uc_mac_addr(), it should use mbx_need_resp flag
to decide whether send response to VF.

Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
---
 drivers/net/ethernet/hisilicon/hns3/hclge_mbx.h          | 5 ++++-
 drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_mbx.c   | 7 +++----
 drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_mbx.c | 2 ++
 3 files changed, 9 insertions(+), 5 deletions(-)

diff --git a/drivers/net/ethernet/hisilicon/hns3/hclge_mbx.h b/drivers/net/ethernet/hisilicon/hns3/hclge_mbx.h
index 8b6191f..83e19c6 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hclge_mbx.h
+++ b/drivers/net/ethernet/hisilicon/hns3/hclge_mbx.h
@@ -84,12 +84,15 @@ struct hclgevf_mbx_resp_status {
 struct hclge_mbx_vf_to_pf_cmd {
 	u8 rsv;
 	u8 mbx_src_vfid; /* Auto filled by IMP */
-	u8 rsv1[2];
+	u8 mbx_need_resp;
+	u8 rsv1[1];
 	u8 msg_len;
 	u8 rsv2[3];
 	u8 msg[HCLGE_MBX_MAX_MSG_SIZE];
 };
 
+#define HCLGE_MBX_NEED_RESP_BIT		BIT(0)
+
 struct hclge_mbx_pf_to_vf_cmd {
 	u8 dest_vfid;
 	u8 rsv[3];
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_mbx.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_mbx.c
index 24386bd..fe48c56 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_mbx.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_mbx.c
@@ -212,8 +212,7 @@ static int hclge_set_vf_promisc_mode(struct hclge_vport *vport,
 }
 
 static int hclge_set_vf_uc_mac_addr(struct hclge_vport *vport,
-				    struct hclge_mbx_vf_to_pf_cmd *mbx_req,
-				    bool gen_resp)
+				    struct hclge_mbx_vf_to_pf_cmd *mbx_req)
 {
 	const u8 *mac_addr = (const u8 *)(&mbx_req->msg[2]);
 	struct hclge_dev *hdev = vport->back;
@@ -249,7 +248,7 @@ static int hclge_set_vf_uc_mac_addr(struct hclge_vport *vport,
 		return -EIO;
 	}
 
-	if (gen_resp)
+	if (mbx_req->mbx_need_resp & HCLGE_MBX_NEED_RESP_BIT)
 		hclge_gen_resp_to_vf(vport, mbx_req, status, NULL, 0);
 
 	return 0;
@@ -597,7 +596,7 @@ void hclge_mbx_handler(struct hclge_dev *hdev)
 					ret);
 			break;
 		case HCLGE_MBX_SET_UNICAST:
-			ret = hclge_set_vf_uc_mac_addr(vport, req, true);
+			ret = hclge_set_vf_uc_mac_addr(vport, req);
 			if (ret)
 				dev_err(&hdev->pdev->dev,
 					"PF fail(%d) to set VF UC MAC Addr\n",
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_mbx.c b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_mbx.c
index 3c22639..30f2e93 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_mbx.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_mbx.c
@@ -98,6 +98,8 @@ int hclgevf_send_mbx_msg(struct hclgevf_dev *hdev, u16 code, u16 subcode,
 	}
 
 	hclgevf_cmd_setup_basic_desc(&desc, HCLGEVF_OPC_MBX_VF_TO_PF, false);
+	req->mbx_need_resp |= need_resp ? HCLGE_MBX_NEED_RESP_BIT :
+					  ~HCLGE_MBX_NEED_RESP_BIT;
 	req->msg[0] = code;
 	req->msg[1] = subcode;
 	memcpy(&req->msg[2], msg_data, msg_len);
-- 
2.7.4

[PATCH V2 net-next 07/12] net: hns3: not reset TQP in the DOWN while VF resetting

[Huazhong Tan]

Since the hardware does not handle mailboxes and the hardware
reset include TQP reset, so it is unnecessary to reset TQP
in the hclgevf_ae_stop() while doing VF reset. Also it is
unnecessary to reset the remaining TQP when one reset fails.

Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
---
 drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c
index 7672cab..6ce5b03 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c
@@ -2050,8 +2050,10 @@ static void hclgevf_ae_stop(struct hnae3_handle *handle)
 
 	set_bit(HCLGEVF_STATE_DOWN, &hdev->state);
 
-	for (i = 0; i < handle->kinfo.num_tqps; i++)
-		hclgevf_reset_tqp(handle, i);
+	if (hdev->reset_type != HNAE3_VF_RESET)
+		for (i = 0; i < handle->kinfo.num_tqps; i++)
+			if (hclgevf_reset_tqp(handle, i))
+				break;
 
 	/* reset tqp stats */
 	hclgevf_reset_tqp_stats(handle);
-- 
2.7.4

[PATCH V2 net-next 08/12] net: hns3: stop schedule reset service while unloading driver

[Huazhong Tan]

This patch uses HCLGE_STATE_REMOVING/HCLGEVF_STATE_REMOVING flag to
indicate that the driver is unloading, and we should stop new coming
reset service to be scheduled, otherwise, reset service will access
some resource which has been freed by unloading.

Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
---
V1->V2: fixes a flag setting error
---
 drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c   | 4 +++-
 drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c | 4 +++-
 drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.h | 1 +
 3 files changed, 7 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
index 4d5568e..ead8308 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
@@ -2175,7 +2175,8 @@ static void hclge_mbx_task_schedule(struct hclge_dev *hdev)
 
 static void hclge_reset_task_schedule(struct hclge_dev *hdev)
 {
-	if (!test_and_set_bit(HCLGE_STATE_RST_SERVICE_SCHED, &hdev->state))
+	if (!test_bit(HCLGE_STATE_REMOVING, &hdev->state) &&
+	    !test_and_set_bit(HCLGE_STATE_RST_SERVICE_SCHED, &hdev->state))
 		schedule_work(&hdev->rst_service_task);
 }
 
@@ -7768,6 +7769,7 @@ static void hclge_state_init(struct hclge_dev *hdev)
 static void hclge_state_uninit(struct hclge_dev *hdev)
 {
 	set_bit(HCLGE_STATE_DOWN, &hdev->state);
+	set_bit(HCLGE_STATE_REMOVING, &hdev->state);
 
 	if (hdev->service_timer.function)
 		del_timer_sync(&hdev->service_timer);
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c
index 6ce5b03..01ab843 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c
@@ -1611,7 +1611,8 @@ static void hclgevf_get_misc_vector(struct hclgevf_dev *hdev)
 
 void hclgevf_reset_task_schedule(struct hclgevf_dev *hdev)
 {
-	if (!test_bit(HCLGEVF_STATE_RST_SERVICE_SCHED, &hdev->state)) {
+	if (!test_bit(HCLGEVF_STATE_RST_SERVICE_SCHED, &hdev->state) &&
+	    !test_bit(HCLGEVF_STATE_REMOVING, &hdev->state)) {
 		set_bit(HCLGEVF_STATE_RST_SERVICE_SCHED, &hdev->state);
 		schedule_work(&hdev->rst_service_task);
 	}
@@ -2122,6 +2123,7 @@ static void hclgevf_state_init(struct hclgevf_dev *hdev)
 static void hclgevf_state_uninit(struct hclgevf_dev *hdev)
 {
 	set_bit(HCLGEVF_STATE_DOWN, &hdev->state);
+	set_bit(HCLGEVF_STATE_REMOVING, &hdev->state);
 
 	if (hdev->keep_alive_timer.function)
 		del_timer_sync(&hdev->keep_alive_timer);
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.h b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.h
index ee3a6cb..0cceda3 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.h
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.h
@@ -130,6 +130,7 @@ enum hclgevf_states {
 	HCLGEVF_STATE_DOWN,
 	HCLGEVF_STATE_DISABLED,
 	HCLGEVF_STATE_IRQ_INITED,
+	HCLGEVF_STATE_REMOVING,
 	/* task states */
 	HCLGEVF_STATE_SERVICE_SCHED,
 	HCLGEVF_STATE_RST_SERVICE_SCHED,
-- 
2.7.4

Re: [PATCH V2 net-next 08/12] net: hns3: stop schedule reset service while unloading driver

[Neil Horman]

On Wed, Apr 24, 2019 at 07:05:27PM +0800, Huazhong Tan wrote:
> This patch uses HCLGE_STATE_REMOVING/HCLGEVF_STATE_REMOVING flag to
> indicate that the driver is unloading, and we should stop new coming
> reset service to be scheduled, otherwise, reset service will access
> some resource which has been freed by unloading.
> 
> Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
> Signed-off-by: Peng Li <lipeng321@huawei.com>
> ---
> V1->V2: fixes a flag setting error
> ---
>  drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c   | 4 +++-
>  drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c | 4 +++-
>  drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.h | 1 +
>  3 files changed, 7 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
> index 4d5568e..ead8308 100644
> --- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
> +++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
> @@ -2175,7 +2175,8 @@ static void hclge_mbx_task_schedule(struct hclge_dev *hdev)
>  
>  static void hclge_reset_task_schedule(struct hclge_dev *hdev)
>  {
> -	if (!test_and_set_bit(HCLGE_STATE_RST_SERVICE_SCHED, &hdev->state))
> +	if (!test_bit(HCLGE_STATE_REMOVING, &hdev->state) &&
> +	    !test_and_set_bit(HCLGE_STATE_RST_SERVICE_SCHED, &hdev->state))
>  		schedule_work(&hdev->rst_service_task);
>  }
>  

In what use case do you need an extra bit for this?  From my read, this work
task only gets scheduled from:
1) Interrupt handlers
2) Its own service task

Based on the fact that you are calling cancel_work_sync(...rst_service_task)
from the pci teardown routine, irqs should all be disabled on your devices
already (meaning interrupts shouldn't schedule it), and cancel_work_sync
guarantees that rearming cant happen from within its own service task.

Neil

Re: [PATCH V2 net-next 08/12] net: hns3: stop schedule reset service while unloading driver

[tanhuazhong]

On 2019/4/24 21:55, Neil Horman wrote:
> On Wed, Apr 24, 2019 at 07:05:27PM +0800, Huazhong Tan wrote:
>> This patch uses HCLGE_STATE_REMOVING/HCLGEVF_STATE_REMOVING flag to
>> indicate that the driver is unloading, and we should stop new coming
>> reset service to be scheduled, otherwise, reset service will access
>> some resource which has been freed by unloading.
>>
>> Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
>> Signed-off-by: Peng Li <lipeng321@huawei.com>
>> ---
>> V1->V2: fixes a flag setting error
>> ---
>>   drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c   | 4 +++-
>>   drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c | 4 +++-
>>   drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.h | 1 +
>>   3 files changed, 7 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
>> index 4d5568e..ead8308 100644
>> --- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
>> +++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
>> @@ -2175,7 +2175,8 @@ static void hclge_mbx_task_schedule(struct hclge_dev *hdev)
>>   
>>   static void hclge_reset_task_schedule(struct hclge_dev *hdev)
>>   {
>> -	if (!test_and_set_bit(HCLGE_STATE_RST_SERVICE_SCHED, &hdev->state))
>> +	if (!test_bit(HCLGE_STATE_REMOVING, &hdev->state) &&
>> +	    !test_and_set_bit(HCLGE_STATE_RST_SERVICE_SCHED, &hdev->state))
>>   		schedule_work(&hdev->rst_service_task);
>>   }
>>   
> 
> In what use case do you need an extra bit for this?  From my read, this work
> task only gets scheduled from:
> 1) Interrupt handlers
> 2) Its own service task
> 
> Based on the fact that you are calling cancel_work_sync(...rst_service_task)
> from the pci teardown routine, irqs should all be disabled on your devices
> already (meaning interrupts shouldn't schedule it), and cancel_work_sync
> guarantees that rearming cant happen from within its own service task.
> 
> Neil
>

Beside these two cases, when the client detects an error and requests a 
reset, it will call hclge_reset_event and schedule the reset work task 
to deal with the request. This may happen after calling 
cancel_work_sync(...rst_service_task) from the pci teardown routine.

Best Regards, Huazhong

> 
> .
> 

Re: [PATCH V2 net-next 08/12] net: hns3: stop schedule reset service while unloading driver

[Neil Horman]

On Thu, Apr 25, 2019 at 02:06:07PM +0800, tanhuazhong wrote:
> 
> 
> On 2019/4/24 21:55, Neil Horman wrote:
> > On Wed, Apr 24, 2019 at 07:05:27PM +0800, Huazhong Tan wrote:
> > > This patch uses HCLGE_STATE_REMOVING/HCLGEVF_STATE_REMOVING flag to
> > > indicate that the driver is unloading, and we should stop new coming
> > > reset service to be scheduled, otherwise, reset service will access
> > > some resource which has been freed by unloading.
> > > 
> > > Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
> > > Signed-off-by: Peng Li <lipeng321@huawei.com>
> > > ---
> > > V1->V2: fixes a flag setting error
> > > ---
> > >   drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c   | 4 +++-
> > >   drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c | 4 +++-
> > >   drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.h | 1 +
> > >   3 files changed, 7 insertions(+), 2 deletions(-)
> > > 
> > > diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
> > > index 4d5568e..ead8308 100644
> > > --- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
> > > +++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
> > > @@ -2175,7 +2175,8 @@ static void hclge_mbx_task_schedule(struct hclge_dev *hdev)
> > >   static void hclge_reset_task_schedule(struct hclge_dev *hdev)
> > >   {
> > > -	if (!test_and_set_bit(HCLGE_STATE_RST_SERVICE_SCHED, &hdev->state))
> > > +	if (!test_bit(HCLGE_STATE_REMOVING, &hdev->state) &&
> > > +	    !test_and_set_bit(HCLGE_STATE_RST_SERVICE_SCHED, &hdev->state))
> > >   		schedule_work(&hdev->rst_service_task);
> > >   }
> > 
> > In what use case do you need an extra bit for this?  From my read, this work
> > task only gets scheduled from:
> > 1) Interrupt handlers
> > 2) Its own service task
> > 
> > Based on the fact that you are calling cancel_work_sync(...rst_service_task)
> > from the pci teardown routine, irqs should all be disabled on your devices
> > already (meaning interrupts shouldn't schedule it), and cancel_work_sync
> > guarantees that rearming cant happen from within its own service task.
> > 
> > Neil
> > 
> 
> Beside these two cases, when the client detects an error and requests a
> reset, it will call hclge_reset_event and schedule the reset work task to
> deal with the request. This may happen after calling
> cancel_work_sync(...rst_service_task) from the pci teardown routine.
> 
> Best Regards, Huazhong
> 
But that is handled from either:
1) hns_roce_v2_msix_interrupt_abn
or
2) hns3_nic_net_timeout
or
3) hns3_slot_reset

You should be protected from (1) because interrupts should be disabled before
you call cancel_work_syn.

You should be protected from (2) because the network interface will have been
brought down first (disabling the watchdog timer for the any interfaces on this
hardware

You should be protected from (3) because you are tearing down the pci device
anyway, and the pci subsystem should be ignoring resets

Neil

> > 
> > .
> > 
> 

Re: [PATCH V2 net-next 08/12] net: hns3: stop schedule reset service while unloading driver

[tanhuazhong]

On 2019/4/25 19:41, Neil Horman wrote:
> On Thu, Apr 25, 2019 at 02:06:07PM +0800, tanhuazhong wrote:
>>
>>
>> On 2019/4/24 21:55, Neil Horman wrote:
>>> On Wed, Apr 24, 2019 at 07:05:27PM +0800, Huazhong Tan wrote:
>>>> This patch uses HCLGE_STATE_REMOVING/HCLGEVF_STATE_REMOVING flag to
>>>> indicate that the driver is unloading, and we should stop new coming
>>>> reset service to be scheduled, otherwise, reset service will access
>>>> some resource which has been freed by unloading.
>>>>
>>>> Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
>>>> Signed-off-by: Peng Li <lipeng321@huawei.com>
>>>> ---
>>>> V1->V2: fixes a flag setting error
>>>> ---
>>>>    drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c   | 4 +++-
>>>>    drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c | 4 +++-
>>>>    drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.h | 1 +
>>>>    3 files changed, 7 insertions(+), 2 deletions(-)
>>>>
>>>> diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
>>>> index 4d5568e..ead8308 100644
>>>> --- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
>>>> +++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
>>>> @@ -2175,7 +2175,8 @@ static void hclge_mbx_task_schedule(struct hclge_dev *hdev)
>>>>    static void hclge_reset_task_schedule(struct hclge_dev *hdev)
>>>>    {
>>>> -	if (!test_and_set_bit(HCLGE_STATE_RST_SERVICE_SCHED, &hdev->state))
>>>> +	if (!test_bit(HCLGE_STATE_REMOVING, &hdev->state) &&
>>>> +	    !test_and_set_bit(HCLGE_STATE_RST_SERVICE_SCHED, &hdev->state))
>>>>    		schedule_work(&hdev->rst_service_task);
>>>>    }
>>>
>>> In what use case do you need an extra bit for this?  From my read, this work
>>> task only gets scheduled from:
>>> 1) Interrupt handlers
>>> 2) Its own service task
>>>
>>> Based on the fact that you are calling cancel_work_sync(...rst_service_task)
>>> from the pci teardown routine, irqs should all be disabled on your devices
>>> already (meaning interrupts shouldn't schedule it), and cancel_work_sync
>>> guarantees that rearming cant happen from within its own service task.
>>>
>>> Neil
>>>
>>
>> Beside these two cases, when the client detects an error and requests a
>> reset, it will call hclge_reset_event and schedule the reset work task to
>> deal with the request. This may happen after calling
>> cancel_work_sync(...rst_service_task) from the pci teardown routine.
>>
>> Best Regards, Huazhong
>>
> But that is handled from either:
> 1) hns_roce_v2_msix_interrupt_abn
> or
> 2) hns3_nic_net_timeout
> or
> 3) hns3_slot_reset
> 
> You should be protected from (1) because interrupts should be disabled before
> you call cancel_work_syn.
> 
> You should be protected from (2) because the network interface will have been
> brought down first (disabling the watchdog timer for the any interfaces on this
> hardware
> 
> You should be protected from (3) because you are tearing down the pci device
> anyway, and the pci subsystem should be ignoring resets
> 
> Neil
> 

Yes, your analysis is right, i will remove this patch.

Thanks, Huazhong.

>>>
>>> .
>>>
>>
> 
> .
> 

[PATCH V2 net-next 09/12] net: hns3: fix pause configure fail problem

[Huazhong Tan]

When configure pause, current implementation returns directly
after setup PFC without setup BP, which is not sufficient.

So this patch fixes it, only return while setting PFC failed.

Fixes: 44e59e375bf7 ("net: hns3: do not return GE PFC setting err when initializing")
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
---
 drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_tm.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_tm.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_tm.c
index aafc69f..a7bbb6d 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_tm.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_tm.c
@@ -1331,8 +1331,11 @@ int hclge_pause_setup_hw(struct hclge_dev *hdev, bool init)
 	ret = hclge_pfc_setup_hw(hdev);
 	if (init && ret == -EOPNOTSUPP)
 		dev_warn(&hdev->pdev->dev, "GE MAC does not support pfc\n");
-	else
+	else if (ret) {
+		dev_err(&hdev->pdev->dev, "config pfc failed! ret = %d\n",
+			ret);
 		return ret;
+	}
 
 	return hclge_tm_bp_setup(hdev);
 }
-- 
2.7.4

[PATCH V2 net-next 10/12] net: hns3: extend the loopback state acquisition time

[Huazhong Tan]

From: liuzhongzhu <liuzhongzhu@huawei.com>

The test results show that the maximum time of hardware return
to mac link state is 500MS.The software needs to set twice the
maximum time of hardware return state (1000MS).

If not modified, the loopback test returns probability failure.

Signed-off-by: liuzhongzhu <liuzhongzhu@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
---
 drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
index ead8308..22afd75 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
@@ -5338,8 +5338,8 @@ static int hclge_set_serdes_loopback(struct hclge_dev *hdev, bool en,
 #define HCLGE_SERDES_RETRY_MS	10
 #define HCLGE_SERDES_RETRY_NUM	100
 
-#define HCLGE_MAC_LINK_STATUS_MS   20
-#define HCLGE_MAC_LINK_STATUS_NUM  10
+#define HCLGE_MAC_LINK_STATUS_MS   10
+#define HCLGE_MAC_LINK_STATUS_NUM  100
 #define HCLGE_MAC_LINK_STATUS_DOWN 0
 #define HCLGE_MAC_LINK_STATUS_UP   1
 
-- 
2.7.4

[PATCH V2 net-next 11/12] net: hns3: prevent double free in hns3_put_ring_config()

[Huazhong Tan]

This patch adds a check for the hns3_put_ring_config() to prevent
double free, and for more readable, move the NULL assignment of
priv->ring_data into the hns3_put_ring_config().

Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
---
 drivers/net/ethernet/hisilicon/hns3/hns3_enet.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
index 7dc281c..b0f31eb 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
@@ -3487,6 +3487,7 @@ static int hns3_get_ring_config(struct hns3_nic_priv *priv)
 	}
 
 	devm_kfree(&pdev->dev, priv->ring_data);
+	priv->ring_data = NULL;
 	return ret;
 }
 
@@ -3495,12 +3496,16 @@ static void hns3_put_ring_config(struct hns3_nic_priv *priv)
 	struct hnae3_handle *h = priv->ae_handle;
 	int i;
 
+	if (!priv->ring_data)
+		return;
+
 	for (i = 0; i < h->kinfo.num_tqps; i++) {
 		devm_kfree(priv->dev, priv->ring_data[i].ring);
 		devm_kfree(priv->dev,
 			   priv->ring_data[i + h->kinfo.num_tqps].ring);
 	}
 	devm_kfree(priv->dev, priv->ring_data);
+	priv->ring_data = NULL;
 }
 
 static int hns3_alloc_ring_memory(struct hns3_enet_ring *ring)
@@ -3920,8 +3925,6 @@ static void hns3_client_uninit(struct hnae3_handle *handle, bool reset)
 
 	hns3_dbg_uninit(handle);
 
-	priv->ring_data = NULL;
-
 out_netdev_free:
 	free_netdev(netdev);
 }
@@ -4268,12 +4271,10 @@ static int hns3_reset_notify_init_enet(struct hnae3_handle *handle)
 	hns3_uninit_all_ring(priv);
 err_uninit_vector:
 	hns3_nic_uninit_vector_data(priv);
-	priv->ring_data = NULL;
 err_dealloc_vector:
 	hns3_nic_dealloc_vector_data(priv);
 err_put_ring:
 	hns3_put_ring_config(priv);
-	priv->ring_data = NULL;
 
 	return ret;
 }
@@ -4335,7 +4336,6 @@ static int hns3_reset_notify_uninit_enet(struct hnae3_handle *handle)
 		netdev_err(netdev, "uninit ring error\n");
 
 	hns3_put_ring_config(priv);
-	priv->ring_data = NULL;
 
 	return ret;
 }
-- 
2.7.4

[PATCH V2 net-next 12/12] net: hns3: remove reset after command send failed

[Huazhong Tan]

From: Weihang Li <liweihang@hisilicon.com>

It's meaningless to trigger reset when failed to send command to IMP,
because the failure is usually caused by no authority, illegal command
and so on. When that happened, we just need to return the status code
for further debugging.

Signed-off-by: Weihang Li <liweihang@hisilicon.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
---
 drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_err.c | 10 ----------
 1 file changed, 10 deletions(-)

diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_err.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_err.c
index 804c870..4ac8063 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_err.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_err.c
@@ -1653,8 +1653,6 @@ int hclge_handle_hw_msix_error(struct hclge_dev *hdev,
 	if (ret) {
 		dev_err(dev, "fail(%d) to query msix int status bd num\n",
 			ret);
-		/* reset everything for now */
-		set_bit(HNAE3_GLOBAL_RESET, reset_requests);
 		return ret;
 	}
 
@@ -1675,8 +1673,6 @@ int hclge_handle_hw_msix_error(struct hclge_dev *hdev,
 	if (ret) {
 		dev_err(dev, "query all mpf msix int cmd failed (%d)\n",
 			ret);
-		/* reset everything for now */
-		set_bit(HNAE3_GLOBAL_RESET, reset_requests);
 		goto msi_error;
 	}
 
@@ -1710,8 +1706,6 @@ int hclge_handle_hw_msix_error(struct hclge_dev *hdev,
 	if (ret) {
 		dev_err(dev, "clear all mpf msix int cmd failed (%d)\n",
 			ret);
-		/* reset everything for now */
-		set_bit(HNAE3_GLOBAL_RESET, reset_requests);
 		goto msi_error;
 	}
 
@@ -1725,8 +1719,6 @@ int hclge_handle_hw_msix_error(struct hclge_dev *hdev,
 	if (ret) {
 		dev_err(dev, "query all pf msix int cmd failed (%d)\n",
 			ret);
-		/* reset everything for now */
-		set_bit(HNAE3_GLOBAL_RESET, reset_requests);
 		goto msi_error;
 	}
 
@@ -1767,8 +1759,6 @@ int hclge_handle_hw_msix_error(struct hclge_dev *hdev,
 	if (ret) {
 		dev_err(dev, "clear all pf msix int cmd failed (%d)\n",
 			ret);
-		/* reset everything for now */
-		set_bit(HNAE3_GLOBAL_RESET, reset_requests);
 	}
 
 	/* query and clear mac tnl interruptions */
-- 
2.7.4