[dpdk-dev] [PATCH v4] net/axgbe: support sfp module EEPROM
Somalapuram, Amaranath
Amaranath.Somalapuram at amd.com
Wed Jan 6 13:31:02 CET 2021
-----Original Message-----
From: Ferruh Yigit <ferruh.yigit at intel.com>
Sent: Monday, January 4, 2021 7:54 PM
To: Somalapuram, Amaranath <Amaranath.Somalapuram at amd.com>; dev at dpdk.org
Subject: Re: [PATCH v4] net/axgbe: support sfp module EEPROM
[CAUTION: External Email]
On 12/22/2020 7:01 AM, asomalap at amd.com wrote:
> From: Amaranath Somalapuram <Amaranath.Somalapuram at amd.com>
>
> Adding API for get_module_eeprom and get_module_info.wq
>
> Signed-off-by: Amaranath Somalapuram <Amaranath.Somalapuram at amd.com>
> ---
> doc/guides/nics/features/axgbe.ini | 1 +
> drivers/net/axgbe/axgbe_ethdev.c | 2 +
> drivers/net/axgbe/axgbe_phy.h | 4 ++
> drivers/net/axgbe/axgbe_phy_impl.c | 108 +++++++++++++++++++++++++++++
> 4 files changed, 115 insertions(+)
>
> diff --git a/doc/guides/nics/features/axgbe.ini
> b/doc/guides/nics/features/axgbe.ini
> index 34df0d1ee..3a090183e 100644
> --- a/doc/guides/nics/features/axgbe.ini
> +++ b/doc/guides/nics/features/axgbe.ini
> @@ -17,6 +17,7 @@ CRC offload = Y
> L3 checksum offload = Y
> L4 checksum offload = Y
> Basic stats = Y
> +Module EEPROM dump = Y
> Linux UIO = Y
> x86-32 = Y
> x86-64 = Y
> diff --git a/drivers/net/axgbe/axgbe_ethdev.c
> b/drivers/net/axgbe/axgbe_ethdev.c
> index cfe6aba73..22e05e8cb 100644
> --- a/drivers/net/axgbe/axgbe_ethdev.c
> +++ b/drivers/net/axgbe/axgbe_ethdev.c
> @@ -257,6 +257,8 @@ static const struct eth_dev_ops axgbe_eth_dev_ops = {
> .timesync_adjust_time = axgbe_timesync_adjust_time,
> .timesync_read_time = axgbe_timesync_read_time,
> .timesync_write_time = axgbe_timesync_write_time,
> + .get_module_info = axgbe_get_module_info,
> + .get_module_eeprom = axgbe_get_module_eeprom,
> };
>
> static int axgbe_phy_reset(struct axgbe_port *pdata) diff --git
> a/drivers/net/axgbe/axgbe_phy.h b/drivers/net/axgbe/axgbe_phy.h index
> 77ee20a31..d9d7fde41 100644
> --- a/drivers/net/axgbe/axgbe_phy.h
> +++ b/drivers/net/axgbe/axgbe_phy.h
> @@ -188,5 +188,9 @@
> #define DUPLEX_FULL 0x01
> #define DUPLEX_UNKNOWN 0xff
>
> +int axgbe_get_module_info(struct rte_eth_dev *dev,
> + struct rte_eth_dev_module_info
> +*modinfo); int axgbe_get_module_eeprom(struct rte_eth_dev *dev,
> + struct rte_dev_eeprom_info *info);
> #endif
> /* PHY */
> diff --git a/drivers/net/axgbe/axgbe_phy_impl.c
> b/drivers/net/axgbe/axgbe_phy_impl.c
> index 02236ec19..af18a8d51 100644
> --- a/drivers/net/axgbe/axgbe_phy_impl.c
> +++ b/drivers/net/axgbe/axgbe_phy_impl.c
> @@ -141,12 +141,18 @@ enum axgbe_sfp_speed {
>
> #define AXGBE_SFP_EXTD_CC 31
>
> +#define AXGBE_SFP_EEPROM_PAGE_SIZE 256
> +
> struct axgbe_sfp_eeprom {
> u8 base[64];
> u8 extd[32];
> u8 vendor[32];
> };
>
> +struct axgbe_sfp_eeprom_module {
> + u8 base[AXGBE_SFP_EEPROM_PAGE_SIZE]; };
> +
> #define AXGBE_BEL_FUSE_VENDOR "BEL-FUSE"
> #define AXGBE_BEL_FUSE_PARTNO "1GBT-SFP06"
>
> @@ -734,6 +740,108 @@ static int axgbe_phy_sfp_read_eeprom(struct axgbe_port *pdata)
> return ret;
> }
>
> +int axgbe_get_module_info(struct rte_eth_dev *dev,
> + struct rte_eth_dev_module_info *modinfo)
> +{
> + struct axgbe_port *pdata = dev->data->dev_private;
> + struct axgbe_sfp_eeprom sfp_eeprom;
> + uint8_t eeprom_addr;
> + int ret;
> +
> + ret = axgbe_phy_get_comm_ownership(pdata);
> +
> + if (ret)
> + return -EIO;
> +
> + ret = axgbe_phy_sfp_get_mux(pdata);
> +
> + if (ret) {
> + PMD_DRV_LOG(ERR, "I2C error setting SFP MUX\n");
> + goto put;
> + }
> +
> + eeprom_addr = 0;
> + ret = axgbe_phy_i2c_read(pdata, AXGBE_SFP_SERIAL_ID_ADDRESS,
> + &eeprom_addr, sizeof(eeprom_addr),
> + &sfp_eeprom, sizeof(sfp_eeprom));
> +
> + if (ret) {
> + PMD_DRV_LOG(ERR, "I2C error reading SFP EEPROM\n");
> + goto put;
> + }
> +
> + if (sfp_eeprom.extd[AXGBE_SFP_EXTD_SFF_8472] != 0xff) {
The 'else' case of this check seems wrong, which returns success but doesn't set modinfo correctly.
I have used the reference code from i40e_ethdev.c and ixgbe_ethdev.c. let me know your inputs.
> + if (sfp_eeprom.extd[AXGBE_SFP_EXTD_SFF_8472] == 0) {
> + modinfo->type = RTE_ETH_MODULE_SFF_8079;
> + modinfo->eeprom_len = RTE_ETH_MODULE_SFF_8079_LEN;
> + } else {
> + modinfo->type = RTE_ETH_MODULE_SFF_8472;
> + modinfo->eeprom_len = RTE_ETH_MODULE_SFF_8472_LEN;
> + }
> + }
> +
> +put:
> + axgbe_phy_sfp_put_mux(pdata);
> + axgbe_phy_put_comm_ownership(pdata);
> + return ret;
> +}
> +
> +int axgbe_get_module_eeprom(struct rte_eth_dev *dev,
> + struct rte_dev_eeprom_info *info) {
> + struct axgbe_port *pdata = dev->data->dev_private;
> + struct axgbe_sfp_eeprom_module sfp_eeprom;
> + uint8_t eeprom_addr;
> + uint8_t *data;
> + uint32_t i;
> + int ret;
> +
> + ret = axgbe_phy_get_comm_ownership(pdata);
> +
> + if (ret)
> + return -EIO;
> +
> + if (!info || !info->length || !info->data) {
Up to you but it may be easier to verify the input and fail before taking the ownership of the device first.
Agreed.
> + axgbe_phy_put_comm_ownership(pdata);
> + return -EINVAL;
> + }
> +
> + ret = axgbe_phy_sfp_get_mux(pdata);
> + if (ret) {
> + PMD_DRV_LOG(ERR, "I2C error setting SFP MUX\n");
> + goto put;
> + }
Just a syntax issue, some lines has an empty line between 'ret' assignment and 'ret' check, but some does not (as above), can you please make it consistent.
Agreed.
> +
> + eeprom_addr = 0;
> + ret = axgbe_phy_i2c_read(pdata, AXGBE_SFP_SERIAL_ID_ADDRESS,
> + &eeprom_addr, sizeof(eeprom_addr),
> + &sfp_eeprom, sizeof(sfp_eeprom));
> + if (ret) {
> + PMD_DRV_LOG(ERR, "I2C error reading SFP EEPROM\n");
> + goto put;
> + }
> + data = info->data;
> +
> + /* for AXGBE_SFP_SERIAL_ID_ADDRESS */
> + for (i = 0; i < AXGBE_SFP_EEPROM_PAGE_SIZE; i++)
> + data[i] = sfp_eeprom.base[i];
Should check if 'info->length' > 'AXGBE_SFP_SERIAL_ID_ADDRESS' before above assignment.
Agreed.
> +
> + eeprom_addr = 0;
> + ret = axgbe_phy_i2c_read(pdata, AXGBE_SFP_DIAG_INFO_ADDRESS,
> + &eeprom_addr, sizeof(eeprom_addr),
> + &sfp_eeprom, sizeof(sfp_eeprom));
> +
> + /* for AXGBE_SFP_DIAG_INFO_ADDRESS */
> + for (i = 0; i < info->length - AXGBE_SFP_EEPROM_PAGE_SIZE; i++)
> + data[i + AXGBE_SFP_EEPROM_PAGE_SIZE] =
> + sfp_eeprom.base[i];
> +
Will it be acceptable if the loop break in the middle because of missing space, "i < info->length - AXGBE_SFP_EEPROM_PAGE_SIZE"? If not, size needs to be checked before the loop.
Agreed.
> +put:
> + axgbe_phy_sfp_put_mux(pdata);
> + axgbe_phy_put_comm_ownership(pdata);
> + return ret;
> +}
> +
> +
> static void axgbe_phy_sfp_signals(struct axgbe_port *pdata)
> {
> struct axgbe_phy_data *phy_data = pdata->phy_data;
>
Please reply for the else case issue. I can resubmit. Other I will fix.
More information about the dev
mailing list