net/i40e: fix X710 Rx issue after reading some registers

  There's an issue that X710 can't receive any packet after
reading some special registers. That's because these
registers are only valid for X722, read access for non-X722
will cause ECC error.

Fixes: d9efd0136ac1 ("i40e: add EEPROM and registers dumping")
Cc: stable@dpdk.org

Signed-off-by: Beilei Xing <beilei.xing@intel.com>
---
 drivers/net/i40e/i40e_ethdev.c | 33 +++++++++++++++++++++++++++++++--
 1 file changed, 31 insertions(+), 2 deletions(-)
  

> -----Original Message-----
> From: Xing, Beilei
> Sent: Tuesday, November 13, 2018 8:11 PM
> To: Zhang, Qi Z <qi.z.zhang@intel.com>; Wu, Jingjing <jingjing.wu@intel.com>
> Cc: dev@dpdk.org; stable@dpdk.org
> Subject: [PATCH] net/i40e: fix X710 Rx issue after reading some registers
> 
> There's an issue that X710 can't receive any packet after reading some special
> registers. That's because these registers are only valid for X722, read access for
> non-X722 will cause ECC error.
> 
> Fixes: d9efd0136ac1 ("i40e: add EEPROM and registers dumping")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Beilei Xing <beilei.xing@intel.com>
> ---
>  drivers/net/i40e/i40e_ethdev.c | 33
> +++++++++++++++++++++++++++++++--
>  1 file changed, 31 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/net/i40e/i40e_ethdev.c b/drivers/net/i40e/i40e_ethdev.c
> index 1c77906..686227e 100644
> --- a/drivers/net/i40e/i40e_ethdev.c
> +++ b/drivers/net/i40e/i40e_ethdev.c
> @@ -11609,6 +11609,31 @@ i40e_dev_rx_queue_intr_disable(struct
> rte_eth_dev *dev, uint16_t queue_id)
>  	return 0;
>  }
> 
> +/**
> + * This function is used to check if the register is valid for non-X722 NIC.
> + * Below is the valid registers list for X722 only:
> + * 0x2b800--0x2bb00
> + * 0x38700--0x38a00
> + * 0x3d800--0x3db00
> + * 0x208e00--0x209000
> + * 0x20be00--0x20c000
> + * 0x263c00--0x264000
> + * 0x265c00--0x266000
> + */
> +static inline int i40e_valid_regs(uint32_t reg_offset) {

The function is named as "i4Oe_xxx", but actually it is for all devices.
I think it will be better either move the mac_type check into the function or rename to a more specific function name.


> +	if (((reg_offset >= 0x2b800) && (reg_offset <= 0x2bb00)) ||
> +	    ((reg_offset >= 0x38700) && (reg_offset <= 0x38a00)) ||
> +	    ((reg_offset >= 0x3d800) && (reg_offset <= 0x3db00)) ||
> +	    ((reg_offset >= 0x208e00) && (reg_offset <= 0x209000)) ||
> +	    ((reg_offset >= 0x20be00) && (reg_offset <= 0x20c000)) ||
> +	    ((reg_offset >= 0x263c00) && (reg_offset <= 0x264000)) ||
> +	    ((reg_offset >= 0x265c00) && (reg_offset <= 0x266000)))
> +		return 0;
> +	else
> +		return 1;
> +}
> +
>  static int i40e_get_regs(struct rte_eth_dev *dev,
>  			 struct rte_dev_reg_info *regs)
>  {
> @@ -11650,8 +11675,12 @@ static int i40e_get_regs(struct rte_eth_dev
> *dev,
>  				reg_offset = arr_idx * reg_info->stride1 +
>  					arr_idx2 * reg_info->stride2;
>  				reg_offset += reg_info->base_addr;
> -				ptr_data[reg_offset >> 2] =
> -					I40E_READ_REG(hw, reg_offset);
> +				if ((hw->mac.type != I40E_MAC_X722) &&
> +				    !i40e_valid_regs(reg_offset))
> +					ptr_data[reg_offset >> 2] = 0;
> +				else
> +					ptr_data[reg_offset >> 2] =
> +						I40E_READ_REG(hw, reg_offset);
>  			}
>  	}
> 
> --
> 2.5.5