DPDK-dev Archive on lore.kernel.org
 help / color / Atom feed
From: "Wei Hu (Xavier)" <huwei013@chinasoftinc.com>
To: Ferruh Yigit <ferruh.yigit@intel.com>
Cc: <dev@dpdk.org>
Subject: Re: [dpdk-dev] [PATCH 2/9] net/hns3: get link state change through mailbox
Date: Sat, 14 Dec 2019 18:26:14 +0800
Message-ID: <d2b62dd1-d54e-4cfc-63aa-fafc8017d8a1@chinasoftinc.com> (raw)
In-Reply-To: <01debb84-b7de-e9fa-10f0-7fbf85dddc17@intel.com>

Hi, Ferruh Yigit

On 2019/12/3 21:16, Ferruh Yigit wrote:
> On 12/2/2019 2:51 AM, Wei Hu (Xavier) wrote:
>> From: Hongbo Zheng <zhenghongbo3@huawei.com>
>>
>> When link down occurs, firmware adds the function of sending message
>> to PF driver through mailbox, hns3 PMD driver can recognize link state
>> change faster through the message.
> 
> Hi Xavier,
> 
> As far as I can see the 'link_update' dev_ops (hns3_dev_link_update()), is just
> copying data from internal structure to "struct rte_eth_link". And you have
> timers set to regularly (ever second?) update the internal structure for link
> status.
> 
> Instead, unless you are not using or need those status internally, you can pull
> the internal link status in the 'hns3_dev_link_update()', this way it can be
> possible to get rid of the timers. Also in current implementation, when used
> asked for the link status, it can be outdated regarding the status of the timers.
> 
We will use the internal link status in the driver, the relevent patch 
will be sent later.

> 
> In this patch, interrupt seems used to update the internal link status, this
> fixes the problem above that user is getting out of date link status. But
> instead of this, what do you think updating "struct rte_eth_link" too in the
> interrupt handler and advertising 'RTE_ETH_DEV_INTR_LSC' capability ("Link
> status event" feature in .ini file), so user can enable the lsc interrupt
> ('dev_conf.intr_conf.lsc') and gets the link status quicker because of the
> support in the API ('rte_eth_link_get')?
> 
Currently, firmware adds the function of sending message to PF driver
through mailbox when the link status is changed, hns3 PMD driver can
usually recognize link state change faster through the message.

And in some extreme cases, this way is not faster than existing method
regularly updating link status by issing the command every second in PF
driver, because of the parallel processing of mailbox and command 
messages in firmware. So we reserve updating link status using timers in 
PF driver and don't advertising 'RTE_ETH_DEV_INTR_LSC' capability.

And we will add querying link status by issuing command to the firmware 
in '.link_update' ops implementation function named hns3_dev_link_update 
to solve the out of date link status in patch V2.

Thanks,
Xavier

> Thanks,
> ferruh
> 
> 
>>
>> Signed-off-by: Hongbo Zheng <zhenghongbo3@huawei.com>
>> Signed-off-by: Wei Hu (Xavier) <xavier.huwei@huawei.com>
>> ---
>>   drivers/net/hns3/hns3_ethdev.c |  8 ++++++--
>>   drivers/net/hns3/hns3_ethdev.h |  1 +
>>   drivers/net/hns3/hns3_mbx.c    | 37 ++++++++++++++++++++++++++++++++++
>>   drivers/net/hns3/hns3_mbx.h    |  8 ++++++++
>>   4 files changed, 52 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/net/hns3/hns3_ethdev.c b/drivers/net/hns3/hns3_ethdev.c
>> index bf0ab458f..3c591be51 100644
>> --- a/drivers/net/hns3/hns3_ethdev.c
>> +++ b/drivers/net/hns3/hns3_ethdev.c
>> @@ -218,6 +218,8 @@ hns3_interrupt_handler(void *param)
>>   		hns3_schedule_reset(hns);
>>   	} else if (event_cause == HNS3_VECTOR0_EVENT_RST)
>>   		hns3_schedule_reset(hns);
>> +	else if (event_cause == HNS3_VECTOR0_EVENT_MBX)
>> +		hns3_dev_handle_mbx_msg(hw);
>>   	else
>>   		hns3_err(hw, "Received unknown event");
>>   
>> @@ -3806,14 +3808,16 @@ hns3_get_mac_link_status(struct hns3_hw *hw)
>>   	return !!link_status;
>>   }
>>   
>> -static void
>> +void
>>   hns3_update_link_status(struct hns3_hw *hw)
>>   {
>>   	int state;
>>   
>>   	state = hns3_get_mac_link_status(hw);
>> -	if (state != hw->mac.link_status)
>> +	if (state != hw->mac.link_status) {
>>   		hw->mac.link_status = state;
>> +		hns3_warn(hw, "Link status change to %s!", state ? "up" : "down");
>> +	}
>>   }
>>   
>>   static void
>> diff --git a/drivers/net/hns3/hns3_ethdev.h b/drivers/net/hns3/hns3_ethdev.h
>> index e9a3fe410..004cd75a9 100644
>> --- a/drivers/net/hns3/hns3_ethdev.h
>> +++ b/drivers/net/hns3/hns3_ethdev.h
>> @@ -631,6 +631,7 @@ int hns3_dev_filter_ctrl(struct rte_eth_dev *dev,
>>   			 enum rte_filter_op filter_op, void *arg);
>>   bool hns3_is_reset_pending(struct hns3_adapter *hns);
>>   bool hns3vf_is_reset_pending(struct hns3_adapter *hns);
>> +void hns3_update_link_status(struct hns3_hw *hw);
>>   
>>   static inline bool
>>   is_reset_pending(struct hns3_adapter *hns)
>> diff --git a/drivers/net/hns3/hns3_mbx.c b/drivers/net/hns3/hns3_mbx.c
>> index c1647af4b..26807bc4b 100644
>> --- a/drivers/net/hns3/hns3_mbx.c
>> +++ b/drivers/net/hns3/hns3_mbx.c
>> @@ -282,6 +282,40 @@ hns3_update_resp_position(struct hns3_hw *hw, uint32_t resp_msg)
>>   	resp->tail = tail;
>>   }
>>   
>> +static void
>> +hns3_link_fail_parse(struct hns3_hw *hw, uint8_t link_fail_code)
>> +{
>> +	switch (link_fail_code) {
>> +	case HNS3_MBX_LF_NORMAL:
>> +		break;
>> +	case HNS3_MBX_LF_REF_CLOCK_LOST:
>> +		hns3_warn(hw, "Reference clock lost!");
>> +		break;
>> +	case HNS3_MBX_LF_XSFP_TX_DISABLE:
>> +		hns3_warn(hw, "SFP tx is disabled!");
>> +		break;
>> +	case HNS3_MBX_LF_XSFP_ABSENT:
>> +		hns3_warn(hw, "SFP is absent!");
>> +		break;
>> +	default:
>> +		hns3_warn(hw, "Unknown fail code:%u!", link_fail_code);
>> +		break;
>> +	}
>> +}
>> +
>> +static void
>> +hns3_handle_link_change_event(struct hns3_hw *hw,
>> +			      struct hns3_mbx_pf_to_vf_cmd *req)
>> +{
>> +#define LINK_STATUS_OFFSET     1
>> +#define LINK_FAIL_CODE_OFFSET  2
>> +
>> +	if (!req->msg[LINK_STATUS_OFFSET])
>> +		hns3_link_fail_parse(hw, req->msg[LINK_FAIL_CODE_OFFSET]);
>> +
>> +	hns3_update_link_status(hw);
>> +}
>> +
>>   void
>>   hns3_dev_handle_mbx_msg(struct hns3_hw *hw)
>>   {
>> @@ -335,6 +369,9 @@ hns3_dev_handle_mbx_msg(struct hns3_hw *hw)
>>   
>>   			hns3_mbx_handler(hw);
>>   			break;
>> +		case HNS3_MBX_PUSH_LINK_STATUS:
>> +			hns3_handle_link_change_event(hw, req);
>> +			break;
>>   		default:
>>   			hns3_err(hw,
>>   				 "VF received unsupported(%d) mbx msg from PF",
>> diff --git a/drivers/net/hns3/hns3_mbx.h b/drivers/net/hns3/hns3_mbx.h
>> index d1a6bfead..3722c8760 100644
>> --- a/drivers/net/hns3/hns3_mbx.h
>> +++ b/drivers/net/hns3/hns3_mbx.h
>> @@ -41,6 +41,7 @@ enum HNS3_MBX_OPCODE {
>>   	HNS3_MBX_GET_QID_IN_PF,         /* (VF -> PF) get queue id in pf */
>>   
>>   	HNS3_MBX_HANDLE_VF_TBL = 38,    /* (VF -> PF) store/clear hw cfg tbl */
>> +	HNS3_MBX_PUSH_LINK_STATUS = 201, /* (IMP -> PF) get port link status */
>>   };
>>   
>>   /* below are per-VF mac-vlan subcodes */
>> @@ -64,6 +65,13 @@ enum hns3_mbx_tbl_cfg_subcode {
>>   	HNS3_MBX_VPORT_LIST_CLEAR = 0,
>>   };
>>   
>> +enum hns3_mbx_link_fail_subcode {
>> +	HNS3_MBX_LF_NORMAL = 0,
>> +	HNS3_MBX_LF_REF_CLOCK_LOST,
>> +	HNS3_MBX_LF_XSFP_TX_DISABLE,
>> +	HNS3_MBX_LF_XSFP_ABSENT,
>> +};
>> +
>>   #define HNS3_MBX_MAX_MSG_SIZE	16
>>   #define HNS3_MBX_MAX_RESP_DATA_SIZE	8
>>   #define HNS3_MBX_RING_MAP_BASIC_MSG_NUM	3
>>

  parent reply index

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-12-02  2:51 [dpdk-dev] [PATCH 0/9] updates for hns3 PMD driver Wei Hu (Xavier)
2019-12-02  2:51 ` [dpdk-dev] [PATCH 1/9] net/hns3: support Rx interrupt Wei Hu (Xavier)
2019-12-02  2:51 ` [dpdk-dev] [PATCH 2/9] net/hns3: get link state change through mailbox Wei Hu (Xavier)
2019-12-03 13:16   ` Ferruh Yigit
2019-12-03 13:19     ` Ferruh Yigit
2019-12-14 10:12       ` Wei Hu (Xavier)
2019-12-14 10:26     ` Wei Hu (Xavier) [this message]
2019-12-02  2:51 ` [dpdk-dev] [PATCH 3/9] net/hns3: modify the return value of enable msix Wei Hu (Xavier)
2019-12-02  2:51 ` [dpdk-dev] [PATCH 4/9] net/hns3: modify custom macro Wei Hu (Xavier)
2019-12-02  2:51 ` [dpdk-dev] [PATCH 5/9] net/hns3: optimize RSS's default algorithm Wei Hu (Xavier)
2019-12-02  2:51 ` [dpdk-dev] [PATCH 6/9] net/hns3: remove the redundant function call Wei Hu (Xavier)
2019-12-02  2:51 ` [dpdk-dev] [PATCH 7/9] net/hns3: remove the redundant variable initialization Wei Hu (Xavier)
2019-12-02  2:51 ` [dpdk-dev] [PATCH 8/9] net/hns3: remove the unnecessary assignment Wei Hu (Xavier)
2019-12-02  2:51 ` [dpdk-dev] [PATCH 9/9] net/hns3: remove the unused macros Wei Hu (Xavier)
  -- strict thread matches above, loose matches on Subject: below --
2019-11-30 10:58 [dpdk-dev] [PATCH 0/9] updates for hns3 PMD driver Wei Hu (Xavier)
2019-11-30 10:58 ` [dpdk-dev] [PATCH 2/9] net/hns3: get link state change through mailbox Wei Hu (Xavier)

Reply instructions:

You may reply publically to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=d2b62dd1-d54e-4cfc-63aa-fafc8017d8a1@chinasoftinc.com \
    --to=huwei013@chinasoftinc.com \
    --cc=dev@dpdk.org \
    --cc=ferruh.yigit@intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

DPDK-dev Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/dpdk-dev/0 dpdk-dev/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 dpdk-dev dpdk-dev/ https://lore.kernel.org/dpdk-dev \
		dev@dpdk.org
	public-inbox-index dpdk-dev

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.dpdk.dev


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git