Re: [dpdk-dev] [PATCH v6 1/2] net/iavf: add IAVF request queues function

["Xu, Ting" <ting.xu@intel.com>]

> -----Original Message-----
> From: Xing, Beilei <beilei.xing@intel.com>
> Sent: Friday, October 16, 2020 4:42 PM
> To: Xu, Ting <ting.xu@intel.com>; dev@dpdk.org
> Cc: Zhang, Qi Z <qi.z.zhang@intel.com>; Wu, Jingjing <jingjing.wu@intel.com>
> Subject: RE: [PATCH v6 1/2] net/iavf: add IAVF request queues function
> 
> 
> 
> > -----Original Message-----
> > From: Xu, Ting <ting.xu@intel.com>
> > Sent: Friday, October 16, 2020 9:43 AM
> > To: dev@dpdk.org
> > Cc: Zhang, Qi Z <qi.z.zhang@intel.com>; Xing, Beilei
> > <beilei.xing@intel.com>; Wu, Jingjing <jingjing.wu@intel.com>; Xu,
> > Ting <ting.xu@intel.com>
> > Subject: [PATCH v6 1/2] net/iavf: add IAVF request queues function
> >
> > Add new virtchnl function to request additional queues from PF.
> > Current default queue pairs number when creating a VF is 16. In order
> > to support up to
> > 256 queue pairs, enable this request queues function.
> > Since request queues command may return event message, modify function
> > iavf_read_msg_from_pf to identify event opcode and mark VF reset status.
> >
> > Signed-off-by: Ting Xu <ting.xu@intel.com>
> > ---
> >  drivers/net/iavf/iavf.h        |   9 ++
> >  drivers/net/iavf/iavf_ethdev.c |  11 +-
> > drivers/net/iavf/iavf_vchnl.c  | 226
> > +++++++++++++++++++++++++--------
> >  3 files changed, 192 insertions(+), 54 deletions(-)
> >
> > diff --git a/drivers/net/iavf/iavf.h b/drivers/net/iavf/iavf.h index
> > d56611608..93c165c62 100644
> > --- a/drivers/net/iavf/iavf.h
> > +++ b/drivers/net/iavf/iavf.h
> > @@ -107,6 +107,14 @@ struct iavf_fdir_info {
> >  /* TODO: is that correct to assume the max number to be 16 ?*/
> >  #define IAVF_MAX_MSIX_VECTORS   16
> >
> > +/* Message type read in admin queue from PF */ enum iavf_aq_result {
> > +	IAVF_MSG_ERR = -1, /* Meet error when accessing admin queue */
> > +	IAVF_MSG_NON,      /* Read nothing from admin queue */
> > +	IAVF_MSG_SYS,      /* Read system msg from admin queue */
> > +	IAVF_MSG_CMD,      /* Read async command result */
> > +};
> 
> Is there no such message type in shared code?
> 

Yes, I did not find similar codes suitable for this case.

> > +
> >  /* Structure to store private data specific for VF instance. */  struct
> iavf_info {
> >  	uint16_t num_queue_pairs;
> > @@ -301,4 +309,5 @@ int iavf_add_del_rss_cfg(struct iavf_adapter
> > *adapter, int iavf_add_del_mc_addr_list(struct iavf_adapter *adapter,
> >  			struct rte_ether_addr *mc_addrs,
> >  			uint32_t mc_addrs_num, bool add);
> > +int iavf_request_queues(struct iavf_adapter *adapter, uint16_t num);
> >  #endif /* _IAVF_ETHDEV_H_ */
> > diff --git a/drivers/net/iavf/iavf_ethdev.c
> > b/drivers/net/iavf/iavf_ethdev.c index 8b1cf8f1c..a4a28b885 100644
> > --- a/drivers/net/iavf/iavf_ethdev.c
> > +++ b/drivers/net/iavf/iavf_ethdev.c
> > @@ -1282,7 +1282,7 @@ iavf_dev_rx_queue_intr_disable(struct
> > rte_eth_dev *dev, uint16_t queue_id)  }
> >
> >  static int
> > -iavf_check_vf_reset_done(struct iavf_hw *hw)
> > +iavf_check_vf_reset_done(struct iavf_hw *hw, struct iavf_info *vf)
> >  {
> >  	int i, reset;
> >
> > @@ -1299,6 +1299,9 @@ iavf_check_vf_reset_done(struct iavf_hw *hw)
> >  	if (i >= IAVF_RESET_WAIT_CNT)
> >  		return -1;
> >
> > +	/* VF is not in reset or reset is completed */
> > +	vf->vf_reset = false;
> 
> Seems it's not related to the feature.
> Is the fix for commit 1eab95fe2e36e191ad85a9aacf82a44e7c8011fc?
> If yes, it's better to separate bug fix from the feature.
> 

I missed the vf->vf_reset setting at the bottom of iavf_dev_init(). It is not needed here.

> > +
> >  	return 0;
> >  }
> >
> > @@ -1666,7 +1669,7 @@ iavf_init_vf(struct rte_eth_dev *dev)
> >  		goto err;
> >  	}
> >
> > -	err = iavf_check_vf_reset_done(hw);
> > +	err = iavf_check_vf_reset_done(hw, vf);
> >  	if (err) {
> >  		PMD_INIT_LOG(ERR, "VF is still resetting");
> >  		goto err;
> > @@ -1911,7 +1914,9 @@ iavf_dev_close(struct rte_eth_dev *dev)
> >
> >  	iavf_dev_stop(dev);
> >  	iavf_flow_flush(dev, NULL);
> > -	iavf_flow_uninit(adapter);
> > +	/* if VF is in reset, adminq is disabled, skip the process via adminq */
> > +	if (!vf->vf_reset)
> > +		iavf_flow_uninit(adapter);
> 
> Same as above.
> 
> >
> >  	/*
> >  	 * disable promiscuous mode before reset vf diff --git
> > a/drivers/net/iavf/iavf_vchnl.c b/drivers/net/iavf/iavf_vchnl.c index
> > 5e7142893..11a1ff608 100644
> > --- a/drivers/net/iavf/iavf_vchnl.c
> > +++ b/drivers/net/iavf/iavf_vchnl.c
> > @@ -17,6 +17,7 @@
> >  #include <rte_eal.h>
> >  #include <rte_ether.h>
> >  #include <rte_ethdev_driver.h>
> > +#include <rte_ethdev_pci.h>
> >  #include <rte_dev.h>
> >
> >  #include "iavf.h"
> > @@ -25,14 +26,54 @@
> >  #define MAX_TRY_TIMES 200
> >  #define ASQ_DELAY_MS  10
> >
> > +static uint32_t
> > +iavf_convert_link_speed(enum virtchnl_link_speed virt_link_speed) {
> > +	uint32_t speed;
> > +
> > +	switch (virt_link_speed) {
> > +	case VIRTCHNL_LINK_SPEED_100MB:
> > +		speed = 100;
> > +		break;
> > +	case VIRTCHNL_LINK_SPEED_1GB:
> > +		speed = 1000;
> > +		break;
> > +	case VIRTCHNL_LINK_SPEED_10GB:
> > +		speed = 10000;
> > +		break;
> > +	case VIRTCHNL_LINK_SPEED_40GB:
> > +		speed = 40000;
> > +		break;
> > +	case VIRTCHNL_LINK_SPEED_20GB:
> > +		speed = 20000;
> > +		break;
> > +	case VIRTCHNL_LINK_SPEED_25GB:
> > +		speed = 25000;
> > +		break;
> > +	case VIRTCHNL_LINK_SPEED_2_5GB:
> > +		speed = 2500;
> > +		break;
> > +	case VIRTCHNL_LINK_SPEED_5GB:
> > +		speed = 5000;
> > +		break;
> > +	default:
> > +		speed = 0;
> > +		break;
> > +	}
> > +
> > +	return speed;
> > +}
> > +
> >  /* Read data in admin queue to get msg from pf driver */ -static enum
> > iavf_status
> > +static enum iavf_aq_result
> >  iavf_read_msg_from_pf(struct iavf_adapter *adapter, uint16_t buf_len,
> >  		     uint8_t *buf)
> >  {
> >  	struct iavf_hw *hw = IAVF_DEV_PRIVATE_TO_HW(adapter);
> >  	struct iavf_info *vf = IAVF_DEV_PRIVATE_TO_VF(adapter);
> > +	struct rte_eth_dev *dev = adapter->eth_dev;
> >  	struct iavf_arq_event_info event;
> > +	enum iavf_aq_result result = IAVF_MSG_NON;
> >  	enum virtchnl_ops opcode;
> >  	int ret;
> >
> > @@ -42,7 +83,9 @@ iavf_read_msg_from_pf(struct iavf_adapter *adapter,
> > uint16_t buf_len,
> >  	/* Can't read any msg from adminQ */
> >  	if (ret) {
> >  		PMD_DRV_LOG(DEBUG, "Can't read msg from AQ");
> > -		return ret;
> > +		if (ret != IAVF_ERR_ADMIN_QUEUE_NO_WORK)
> > +			result = IAVF_MSG_ERR;
> > +		return result;
> >  	}
> >
> >  	opcode = (enum
> > virtchnl_ops)rte_le_to_cpu_32(event.desc.cookie_high);
> > @@ -52,16 +95,51 @@ iavf_read_msg_from_pf(struct iavf_adapter
> > *adapter, uint16_t buf_len,
> >  	PMD_DRV_LOG(DEBUG, "AQ from pf carries opcode %u, retval %d",
> >  		    opcode, vf->cmd_retval);
> >
> > -	if (opcode != vf->pend_cmd) {
> > -		if (opcode != VIRTCHNL_OP_EVENT) {
> > -			PMD_DRV_LOG(WARNING,
> > -				    "command mismatch, expect %u, get %u",
> > -				    vf->pend_cmd, opcode);
> > +	if (opcode == VIRTCHNL_OP_EVENT) {
> > +		struct virtchnl_pf_event *vpe =
> > +			(struct virtchnl_pf_event *)event.msg_buf;
> > +
> > +		result = IAVF_MSG_SYS;
> > +		switch (vpe->event) {
> > +		case VIRTCHNL_EVENT_LINK_CHANGE:
> > +			vf->link_up =
> > +				vpe->event_data.link_event.link_status;
> > +			if (vf->vf_res->vf_cap_flags &
> > +				VIRTCHNL_VF_CAP_ADV_LINK_SPEED) {
> > +				vf->link_speed =
> > +				    vpe-
> > >event_data.link_event_adv.link_speed;
> > +			} else {
> > +				enum virtchnl_link_speed speed;
> > +				speed = vpe-
> > >event_data.link_event.link_speed;
> > +				vf->link_speed =
> > iavf_convert_link_speed(speed);
> > +			}
> > +			iavf_dev_link_update(dev, 0);
> > +			PMD_DRV_LOG(INFO, "Link status update:%s",
> > +					vf->link_up ? "up" : "down");
> > +			break;
> > +		case VIRTCHNL_EVENT_RESET_IMPENDING:
> > +			vf->vf_reset = true;
> > +			PMD_DRV_LOG(INFO, "VF is resetting");
> > +			break;
> > +		case VIRTCHNL_EVENT_PF_DRIVER_CLOSE:
> > +			vf->dev_closed = true;
> > +			PMD_DRV_LOG(INFO, "PF driver closed");
> > +			break;
> > +		default:
> > +			PMD_DRV_LOG(ERR, "%s: Unknown event %d from pf",
> > +					__func__, vpe->event);
> > +		}
> > +	}  else {
> > +		/* async reply msg on command issued by vf previously */
> > +		result = IAVF_MSG_CMD;
> > +		if (opcode != vf->pend_cmd) {
> > +			PMD_DRV_LOG(WARNING, "command mismatch,
> > expect %u, get %u",
> > +					vf->pend_cmd, opcode);
> > +			result = IAVF_MSG_ERR;
> >  		}
> > -		return IAVF_ERR_OPCODE_MISMATCH;
> >  	}
> >
> > -	return IAVF_SUCCESS;
> > +	return result;
> >  }
> 
> How about separate this part which is handling the msg from PF?
> 

Done

> >
> >  static int
> > @@ -69,6 +147,7 @@ iavf_execute_vf_cmd(struct iavf_adapter *adapter,
> > struct iavf_cmd_info *args)  {
> >  	struct iavf_hw *hw = IAVF_DEV_PRIVATE_TO_HW(adapter);
> >  	struct iavf_info *vf = IAVF_DEV_PRIVATE_TO_VF(adapter);
> > +	enum iavf_aq_result result;
> >  	enum iavf_status ret;
> >  	int err = 0;
> >  	int i = 0;
> > @@ -97,9 +176,9 @@ iavf_execute_vf_cmd(struct iavf_adapter *adapter,
> > struct iavf_cmd_info *args)
> >  	case VIRTCHNL_OP_GET_SUPPORTED_RXDIDS:
> >  		/* for init virtchnl ops, need to poll the response */
> >  		do {
> > -			ret = iavf_read_msg_from_pf(adapter, args->out_size,
> > +			result = iavf_read_msg_from_pf(adapter, args-
> > >out_size,
> >  						   args->out_buffer);
> > -			if (ret == IAVF_SUCCESS)
> > +			if (result == IAVF_MSG_CMD)
> >  				break;
> >  			rte_delay_ms(ASQ_DELAY_MS);
> >  		} while (i++ < MAX_TRY_TIMES);
> > @@ -111,7 +190,33 @@ iavf_execute_vf_cmd(struct iavf_adapter *adapter,
> > struct iavf_cmd_info *args)
> >  		}
> >  		_clear_cmd(vf);
> >  		break;
> > -
> > +	case VIRTCHNL_OP_REQUEST_QUEUES:
> > +		/*
> > +		 * ignore async reply, only wait for system message,
> > +		 * vf_reset = true if get VIRTCHNL_EVENT_RESET_IMPENDING,
> > +		 * if not, means request queues failed.
> > +		 */
> > +		do {
> > +			result = iavf_read_msg_from_pf(adapter, args-
> > >out_size,
> > +						   args->out_buffer);
> > +			if (result == IAVF_MSG_SYS && vf->vf_reset) {
> > +				break;
> > +			} else if (result == IAVF_MSG_CMD ||
> > +				result == IAVF_MSG_ERR) {
> > +				err = -1;
> > +				break;
> > +			}
> > +			rte_delay_ms(ASQ_DELAY_MS);
> > +			/* If don't read msg or read sys event, continue */
> > +		} while (i++ < MAX_TRY_TIMES);
> > +		if (i >= MAX_TRY_TIMES ||
> > +		    vf->cmd_retval != VIRTCHNL_STATUS_SUCCESS) {
> > +			err = -1;
> > +			PMD_DRV_LOG(ERR, "No response or return failure
> > (%d)"
> > +				    " for cmd %d", vf->cmd_retval, args->ops);
> > +		}
> > +		_clear_cmd(vf);
> > +		break;
> >  	default:
> >  		/* For other virtchnl ops in running time,
> >  		 * wait for the cmd done flag.
> > @@ -136,44 +241,6 @@ iavf_execute_vf_cmd(struct iavf_adapter *adapter,
> > struct iavf_cmd_info *args)
> >  	return err;
> >  }
> >
> > -static uint32_t
> > -iavf_convert_link_speed(enum virtchnl_link_speed virt_link_speed) -{
> > -	uint32_t speed;
> > -
> > -	switch (virt_link_speed) {
> > -	case VIRTCHNL_LINK_SPEED_100MB:
> > -		speed = 100;
> > -		break;
> > -	case VIRTCHNL_LINK_SPEED_1GB:
> > -		speed = 1000;
> > -		break;
> > -	case VIRTCHNL_LINK_SPEED_10GB:
> > -		speed = 10000;
> > -		break;
> > -	case VIRTCHNL_LINK_SPEED_40GB:
> > -		speed = 40000;
> > -		break;
> > -	case VIRTCHNL_LINK_SPEED_20GB:
> > -		speed = 20000;
> > -		break;
> > -	case VIRTCHNL_LINK_SPEED_25GB:
> > -		speed = 25000;
> > -		break;
> > -	case VIRTCHNL_LINK_SPEED_2_5GB:
> > -		speed = 2500;
> > -		break;
> > -	case VIRTCHNL_LINK_SPEED_5GB:
> > -		speed = 5000;
> > -		break;
> > -	default:
> > -		speed = 0;
> > -		break;
> > -	}
> > -
> > -	return speed;
> > -}
> > -
> >  static void
> >  iavf_handle_pf_event_msg(struct rte_eth_dev *dev, uint8_t *msg,
> >  			uint16_t msglen)
> > @@ -389,7 +456,8 @@ iavf_get_vf_resource(struct iavf_adapter *adapter)
> >  	caps = IAVF_BASIC_OFFLOAD_CAPS |
> > VIRTCHNL_VF_CAP_ADV_LINK_SPEED |
> >  		VIRTCHNL_VF_OFFLOAD_RX_FLEX_DESC |
> >  		VIRTCHNL_VF_OFFLOAD_FDIR_PF |
> > -		VIRTCHNL_VF_OFFLOAD_ADV_RSS_PF;
> > +		VIRTCHNL_VF_OFFLOAD_ADV_RSS_PF |
> > +		VIRTCHNL_VF_OFFLOAD_REQ_QUEUES;
> >
> >  	args.in_args = (uint8_t *)&caps;
> >  	args.in_args_size = sizeof(caps);
> > @@ -1148,3 +1216,59 @@ iavf_add_del_mc_addr_list(struct iavf_adapter
> > *adapter,
> >
> >  	return 0;
> >  }
> > +
> > +int
> > +iavf_request_queues(struct iavf_adapter *adapter, uint16_t num) {
> > +	struct rte_eth_dev *dev = adapter->eth_dev;
> > +	struct iavf_info *vf =  IAVF_DEV_PRIVATE_TO_VF(adapter);
> > +	struct rte_pci_device *pci_dev = RTE_ETH_DEV_TO_PCI(dev);
> > +	struct virtchnl_vf_res_request vfres;
> > +	struct iavf_cmd_info args;
> > +	uint16_t num_queue_pairs;
> > +	int err;
> > +
> > +	if (!(vf->vf_res->vf_cap_flags &
> > +		VIRTCHNL_VF_OFFLOAD_REQ_QUEUES)) {
> > +		PMD_DRV_LOG(ERR, "request queues not supported");
> > +		return -1;
> > +	}
> > +
> > +	if (num == 0) {
> > +		PMD_DRV_LOG(ERR, "queue number cannot be zero");
> > +		return -1;
> > +	}
> > +	vfres.num_queue_pairs = num;
> > +
> > +	args.ops = VIRTCHNL_OP_REQUEST_QUEUES;
> > +	args.in_args = (u8 *)&vfres;
> > +	args.in_args_size = sizeof(vfres);
> > +	args.out_buffer = vf->aq_resp;
> > +	args.out_size = IAVF_AQ_BUF_SZ;
> > +
> > +	/*
> > +	 * disable interrupt to avoid the admin queue message to be read
> > +	 * before iavf_read_msg_from_pf.
> > +	 */
> > +	rte_intr_disable(&pci_dev->intr_handle);
> > +	err = iavf_execute_vf_cmd(adapter, &args);
> > +	rte_intr_enable(&pci_dev->intr_handle);
> > +	if (err) {
> > +		PMD_DRV_LOG(ERR, "fail to execute command
> > OP_REQUEST_QUEUES");
> > +		return err;
> > +	}
> > +
> > +	/* request queues succeeded, vf is resetting */
> > +	if (vf->vf_reset) {
> > +		PMD_DRV_LOG(INFO, "vf is resetting");
> > +		return 0;
> > +	}
> > +
> > +	/* request additional queues failed, return available number */
> > +	num_queue_pairs =
> > +	  ((struct virtchnl_vf_res_request *)args.out_buffer)-
> > >num_queue_pairs;
> > +	PMD_DRV_LOG(ERR, "request queues failed, only %u queues "
> > +		"available", num_queue_pairs);
> > +
> > +	return -1;
> > +}
> > --
> > 2.17.1