This patch-set presents some fixes and minor changes to the HNS3 Ethernet Driver.
Fuyun Liang (2):
net: hns3: Fix for using wrong mask and shift in
hclge_get_ring_chain_from_mbx
net: hns3: Fix comments for hclge_get_ring_chain_from_mbx
Huazhong Tan (6):
net: hns3: Modify the order of initializing command queue register
net: hns3: Reset net device with rtnl_lock
net: hns3: Prevent to request reset frequently
net: hns3: Correct reset event status register
net: hns3: Fix return value error in hns3_reset_notify_down_enet
net: hns3: remove unnecessary ring configuration operation while
resetting
Yunsheng Lin (1):
net: hns3: Fix for reset_level default assignment probelm
drivers/net/ethernet/hisilicon/hns3/hns3_enet.c | 16 ++++++----------
.../net/ethernet/hisilicon/hns3/hns3pf/hclge_cmd.c | 4 ++--
.../net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c | 20 +++++++++++++-------
.../net/ethernet/hisilicon/hns3/hns3pf/hclge_main.h | 1 +
.../net/ethernet/hisilicon/hns3/hns3pf/hclge_mbx.c | 14 ++++++++------
5 files changed, 30 insertions(+), 25 deletions(-)
--
2.7.4
From: Huazhong Tan <[email protected]>
Netdevice reset should not be requested frequently, a new one
must wait a moment since there may be some work not completed.
Signed-off-by: Huazhong Tan <[email protected]>
Signed-off-by: Peng Li <[email protected]>
Signed-off-by: Salil Mehta <[email protected]>
---
drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c | 13 ++++++++++---
1 file changed, 10 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
index 3886290..a1886a3 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
@@ -2819,7 +2819,10 @@ static void hclge_clear_reset_cause(struct hclge_dev *hdev)
static void hclge_reset(struct hclge_dev *hdev)
{
+ struct hnae3_handle *handle;
+
/* perform reset of the stack & ae device for a client */
+ handle = &hdev->vport[0].nic;
rtnl_lock();
hclge_notify_client(hdev, HNAE3_DOWN_CLIENT);
@@ -2836,6 +2839,7 @@ static void hclge_reset(struct hclge_dev *hdev)
}
hclge_notify_client(hdev, HNAE3_UP_CLIENT);
+ handle->last_reset_time = jiffies;
rtnl_unlock();
}
@@ -2849,8 +2853,13 @@ static void hclge_reset_event(struct hnae3_handle *handle)
* know this if last reset request did not occur very recently (watchdog
* timer = 5*HZ, let us check after sufficiently large time, say 4*5*Hz)
* In case of new request we reset the "reset level" to PF reset.
+ * And if it is a repeat reset request of the most recent one then we
+ * want to make sure we throttle the reset request. Therefore, we will
+ * not allow it again before 3*HZ times.
*/
- if (time_after(jiffies, (handle->last_reset_time + 4 * 5 * HZ)))
+ if (time_before(jiffies, (handle->last_reset_time + 3 * HZ)))
+ return;
+ else if (time_after(jiffies, (handle->last_reset_time + 4 * 5 * HZ)))
handle->reset_level = HNAE3_FUNC_RESET;
dev_info(&hdev->pdev->dev, "received reset event , reset type is %d",
@@ -2862,8 +2871,6 @@ static void hclge_reset_event(struct hnae3_handle *handle)
if (handle->reset_level < HNAE3_GLOBAL_RESET)
handle->reset_level++;
-
- handle->last_reset_time = jiffies;
}
static void hclge_reset_subtask(struct hclge_dev *hdev)
--
2.7.4
From: Huazhong Tan <[email protected]>
Since current locking was not covering certain code where
netdev was being accessed or manipulated, this patch fixes
it.
Signed-off-by: Huazhong Tan <[email protected]>
Signed-off-by: Peng Li <[email protected]>
Signed-off-by: Salil Mehta <[email protected]>
---
drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
index 8bbf4e5..3886290 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
@@ -2820,15 +2820,13 @@ static void hclge_clear_reset_cause(struct hclge_dev *hdev)
static void hclge_reset(struct hclge_dev *hdev)
{
/* perform reset of the stack & ae device for a client */
-
+ rtnl_lock();
hclge_notify_client(hdev, HNAE3_DOWN_CLIENT);
if (!hclge_reset_wait(hdev)) {
- rtnl_lock();
hclge_notify_client(hdev, HNAE3_UNINIT_CLIENT);
hclge_reset_ae_dev(hdev->ae_dev);
hclge_notify_client(hdev, HNAE3_INIT_CLIENT);
- rtnl_unlock();
hclge_clear_reset_cause(hdev);
} else {
@@ -2838,6 +2836,7 @@ static void hclge_reset(struct hclge_dev *hdev)
}
hclge_notify_client(hdev, HNAE3_UP_CLIENT);
+ rtnl_unlock();
}
static void hclge_reset_event(struct hnae3_handle *handle)
--
2.7.4
From: Fuyun Liang <[email protected]>
Actually, hclge_get_ring_chain_from_mbx is used to get ring type, tqp id,
and int_gl index from mailbox message. So the comments is incorrect. This
patch fixes it.
Fixes: dde1a86e93ca ("net: hns3: Add mailbox support to PF driver")
Signed-off-by: Fuyun Liang <[email protected]>
Signed-off-by: Peng Li <[email protected]>
Signed-off-by: Salil Mehta <[email protected]>
---
drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_mbx.c | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_mbx.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_mbx.c
index 9d36bcc..f34851c 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_mbx.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_mbx.c
@@ -104,13 +104,15 @@ static void hclge_free_vector_ring_chain(struct hnae3_ring_chain_node *head)
}
}
-/* hclge_get_ring_chain_from_mbx: get ring type & tqpid from mailbox message
+/* hclge_get_ring_chain_from_mbx: get ring type & tqp id & int_gl idx
+ * from mailbox message
* msg[0]: opcode
* msg[1]: <not relevant to this function>
* msg[2]: ring_num
* msg[3]: first ring type (TX|RX)
* msg[4]: first tqp id
- * msg[5] ~ msg[14]: other ring type and tqp id
+ * msg[5]: first int_gl idx
+ * msg[6] ~ msg[14]: other ring type, tqp id and int_gl idx
*/
static int hclge_get_ring_chain_from_mbx(
struct hclge_mbx_vf_to_pf_cmd *req,
--
2.7.4
From: Huazhong Tan <[email protected]>
When doing reset, netdev has not been brought up is not an error,
it means that we do not need do the stop operation, so just return
zero.
Fixes: 76ad4f0ee747 ("net: hns3: Add support of HNS3 Ethernet Driver for hip08 SoC")
Signed-off-by: Huazhong Tan <[email protected]>
Signed-off-by: Peng Li <[email protected]>
Signed-off-by: Salil Mehta <[email protected]>
---
drivers/net/ethernet/hisilicon/hns3/hns3_enet.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
index c211450..8296446 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
@@ -3396,7 +3396,7 @@ static int hns3_reset_notify_down_enet(struct hnae3_handle *handle)
struct net_device *ndev = kinfo->netdev;
if (!netif_running(ndev))
- return -EIO;
+ return 0;
return hns3_nic_net_stop(ndev);
}
--
2.7.4
From: Yunsheng Lin <[email protected]>
handle->reset_level is assigned to HNAE3_NONE_RESET when client is
initialized, if a tx timeout happens right after initialization,
then handle->reset_level is not resetted to HNAE3_FUNC_RESET in
hclge_reset_event, which will cause reset event not properly
handled problem.
This patch fixes it by setting handle->reset_level properly when
client is initialized.
Fixes: 6d4c3981a8d8 ("net: hns3: Changes to make enet watchdog timeout func common for PF/VF")
Signed-off-by: Yunsheng Lin <[email protected]>
Signed-off-by: Peng Li <[email protected]>
Signed-off-by: Salil Mehta <[email protected]>
---
drivers/net/ethernet/hisilicon/hns3/hns3_enet.c | 6 +++++-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
index 24f82b7..29be96e 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
@@ -3085,7 +3085,6 @@ static int hns3_client_init(struct hnae3_handle *handle)
priv->dev = &pdev->dev;
priv->netdev = netdev;
priv->ae_handle = handle;
- priv->ae_handle->reset_level = HNAE3_NONE_RESET;
priv->ae_handle->last_reset_time = jiffies;
priv->tx_timeout_count = 0;
@@ -3106,6 +3105,11 @@ static int hns3_client_init(struct hnae3_handle *handle)
/* Carrier off reporting is important to ethtool even BEFORE open */
netif_carrier_off(netdev);
+ if (handle->flags & HNAE3_SUPPORT_VF)
+ handle->reset_level = HNAE3_VF_RESET;
+ else
+ handle->reset_level = HNAE3_FUNC_RESET;
+
ret = hns3_get_ring_config(priv);
if (ret) {
ret = -ENOMEM;
--
2.7.4
From: Fuyun Liang <[email protected]>
HCLGE_INT_GL_IDX_M and HCLGE_INT_GL_IDX_S are used to set fireware
cmd. When getting int_gl value from mailbox message, we should use
HNAE3_RING_GL_IDX_M and HNAE3_RING_GL_IDX_S.
Fixes: 79eee4108541 ("net: hns3: add int_gl_idx setup for VF")
Signed-off-by: Fuyun Liang <[email protected]>
Signed-off-by: Peng Li <[email protected]>
Signed-off-by: Salil Mehta <[email protected]>
---
drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_mbx.c | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_mbx.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_mbx.c
index 50ae2f8..9d36bcc 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_mbx.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_mbx.c
@@ -131,8 +131,8 @@ static int hclge_get_ring_chain_from_mbx(
hnae3_set_bit(ring_chain->flag, HNAE3_RING_TYPE_B, req->msg[3]);
ring_chain->tqp_index =
hclge_get_queue_id(vport->nic.kinfo.tqp[req->msg[4]]);
- hnae3_set_field(ring_chain->int_gl_idx, HCLGE_INT_GL_IDX_M,
- HCLGE_INT_GL_IDX_S,
+ hnae3_set_field(ring_chain->int_gl_idx, HNAE3_RING_GL_IDX_M,
+ HNAE3_RING_GL_IDX_S,
req->msg[5]);
cur_chain = ring_chain;
@@ -151,8 +151,8 @@ static int hclge_get_ring_chain_from_mbx(
[req->msg[HCLGE_MBX_RING_NODE_VARIABLE_NUM * i +
HCLGE_MBX_RING_MAP_BASIC_MSG_NUM + 1]]);
- hnae3_set_field(new_chain->int_gl_idx, HCLGE_INT_GL_IDX_M,
- HCLGE_INT_GL_IDX_S,
+ hnae3_set_field(new_chain->int_gl_idx, HNAE3_RING_GL_IDX_M,
+ HNAE3_RING_GL_IDX_S,
req->msg[HCLGE_MBX_RING_NODE_VARIABLE_NUM * i +
HCLGE_MBX_RING_MAP_BASIC_MSG_NUM + 2]);
--
2.7.4
From: Huazhong Tan <[email protected]>
The configuration of the ring will be used to reinitialize the
ring after the hardware reset is completed. So we should not
release and reacquire this configuration during reset.
Fixes: bb6b94a896d4 ("net: hns3: Add reset interface implementation in client")
Signed-off-by: Huazhong Tan <[email protected]>
Signed-off-by: Peng Li <[email protected]>
Signed-off-by: Salil Mehta <[email protected]>
---
drivers/net/ethernet/hisilicon/hns3/hns3_enet.c | 8 --------
1 file changed, 8 deletions(-)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
index 8296446..24f82b7 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
@@ -3436,10 +3436,6 @@ static int hns3_reset_notify_init_enet(struct hnae3_handle *handle)
/* Carrier off reporting is important to ethtool even BEFORE open */
netif_carrier_off(netdev);
- ret = hns3_get_ring_config(priv);
- if (ret)
- return ret;
-
ret = hns3_nic_init_vector_data(priv);
if (ret)
return ret;
@@ -3471,10 +3467,6 @@ static int hns3_reset_notify_uninit_enet(struct hnae3_handle *handle)
if (ret)
netdev_err(netdev, "uninit ring error\n");
- hns3_put_ring_config(priv);
-
- priv->ring_data = NULL;
-
hns3_uninit_mac_addr(netdev);
return ret;
--
2.7.4
From: Huazhong Tan <[email protected]>
According to hardware's description, the head pointer register should
be written before the tail pointer register while doing command queue
initialization.
Signed-off-by: Huazhong Tan <[email protected]>
Signed-off-by: Peng Li <[email protected]>
Signed-off-by: Salil Mehta <[email protected]>
---
drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_cmd.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_cmd.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_cmd.c
index eca4b23..cf40afc 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_cmd.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_cmd.c
@@ -119,8 +119,8 @@ static void hclge_cmd_config_regs(struct hclge_cmq_ring *ring)
hclge_write_dev(hw, HCLGE_NIC_CSQ_DEPTH_REG,
(ring->desc_num >> HCLGE_NIC_CMQ_DESC_NUM_S) |
HCLGE_NIC_CMQ_ENABLE);
- hclge_write_dev(hw, HCLGE_NIC_CSQ_TAIL_REG, 0);
hclge_write_dev(hw, HCLGE_NIC_CSQ_HEAD_REG, 0);
+ hclge_write_dev(hw, HCLGE_NIC_CSQ_TAIL_REG, 0);
} else {
hclge_write_dev(hw, HCLGE_NIC_CRQ_BASEADDR_L_REG,
lower_32_bits(dma));
@@ -129,8 +129,8 @@ static void hclge_cmd_config_regs(struct hclge_cmq_ring *ring)
hclge_write_dev(hw, HCLGE_NIC_CRQ_DEPTH_REG,
(ring->desc_num >> HCLGE_NIC_CMQ_DESC_NUM_S) |
HCLGE_NIC_CMQ_ENABLE);
- hclge_write_dev(hw, HCLGE_NIC_CRQ_TAIL_REG, 0);
hclge_write_dev(hw, HCLGE_NIC_CRQ_HEAD_REG, 0);
+ hclge_write_dev(hw, HCLGE_NIC_CRQ_TAIL_REG, 0);
}
}
--
2.7.4
From: Huazhong Tan <[email protected]>
According to hardware's description, driver should get reset event
from VECTOR0_PF_OTHER_INT_ST(0x20800) instead of
VECTOR0_PF_OTHER_INT_SRC(0x20700).
Signed-off-by: Huazhong Tan <[email protected]>
Signed-off-by: Peng Li <[email protected]>
Signed-off-by: Salil Mehta <[email protected]>
---
drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c | 2 +-
drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.h | 1 +
2 files changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
index a1886a3..266c686 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
@@ -2495,7 +2495,7 @@ static u32 hclge_check_event_cause(struct hclge_dev *hdev, u32 *clearval)
u32 cmdq_src_reg;
/* fetch the events from their corresponding regs */
- rst_src_reg = hclge_read_dev(&hdev->hw, HCLGE_MISC_RESET_STS_REG);
+ rst_src_reg = hclge_read_dev(&hdev->hw, HCLGE_MISC_VECTOR_INT_STS);
cmdq_src_reg = hclge_read_dev(&hdev->hw, HCLGE_VECTOR0_CMDQ_SRC_REG);
/* Assumption: If by any chance reset and mailbox events are reported
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.h b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.h
index 20abe82..a5abf8e 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.h
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.h
@@ -89,6 +89,7 @@
/* Reset related Registers */
#define HCLGE_MISC_RESET_STS_REG 0x20700
+#define HCLGE_MISC_VECTOR_INT_STS 0x20800
#define HCLGE_GLOBAL_RESET_REG 0x20A00
#define HCLGE_GLOBAL_RESET_BIT 0x0
#define HCLGE_CORE_RESET_BIT 0x1
--
2.7.4
From: Salil Mehta <[email protected]>
Date: Mon, 16 Jul 2018 16:36:18 +0100
> This patch-set presents some fixes and minor changes to the HNS3
> Ethernet Driver.
Series applied, thanks.
Hello!
On 7/16/2018 6:36 PM, Salil Mehta wrote:
> From: Yunsheng Lin <[email protected]>
>
> handle->reset_level is assigned to HNAE3_NONE_RESET when client is
> initialized, if a tx timeout happens right after initialization,
> then handle->reset_level is not resetted to HNAE3_FUNC_RESET in
s/resetted/reset/.
> hclge_reset_event, which will cause reset event not properly
> handled problem.
>
> This patch fixes it by setting handle->reset_level properly when
> client is initialized.
>
> Fixes: 6d4c3981a8d8 ("net: hns3: Changes to make enet watchdog timeout func common for PF/VF")
> Signed-off-by: Yunsheng Lin <[email protected]>
> Signed-off-by: Peng Li <[email protected]>
> Signed-off-by: Salil Mehta <[email protected]>
[...]
MBR, Sergei
Hi Sergei,
> -----Original Message-----
> From: [email protected] [mailto:netdev-
> [email protected]] On Behalf Of Sergei Shtylyov
> Sent: Tuesday, July 17, 2018 10:37 AM
> To: Salil Mehta <[email protected]>; [email protected]
> Cc: Zhuangyuzeng (Yisen) <[email protected]>; lipeng (Y)
> <[email protected]>; [email protected]; [email protected];
> [email protected]; Linuxarm <[email protected]>;
> linyunsheng <[email protected]>
> Subject: Re: [PATCH net-next 7/9] net: hns3: Fix for reset_level
> default assignment probelm
>
> Hello!
>
> On 7/16/2018 6:36 PM, Salil Mehta wrote:
>
> > From: Yunsheng Lin <[email protected]>
> >
> > handle->reset_level is assigned to HNAE3_NONE_RESET when client is
> > initialized, if a tx timeout happens right after initialization,
> > then handle->reset_level is not resetted to HNAE3_FUNC_RESET in
>
> s/resetted/reset/.
Thanks for looking through. Will take care in future submissions.
Best regards
Salil
>
> > hclge_reset_event, which will cause reset event not properly
> > handled problem.
> >
> > This patch fixes it by setting handle->reset_level properly when
> > client is initialized.
> >
> > Fixes: 6d4c3981a8d8 ("net: hns3: Changes to make enet watchdog
> timeout func common for PF/VF")
> > Signed-off-by: Yunsheng Lin <[email protected]>
> > Signed-off-by: Peng Li <[email protected]>
> > Signed-off-by: Salil Mehta <[email protected]>
> [...]
>
> MBR, Sergei