On 16/04/2019 15:34, Heiner Kallweit wrote:
On 17.04.2019 00:25, Andrew Lunn wrote:
@@ -194,7 +191,7 @@ static struct phy_driver bcm87xx_driver[] = {
.phy_id = PHY_ID_BCM8706,
.phy_id_mask = 0xffffffff,
.name = "Broadcom BCM8706",
- .features = PHY_10GBIT_FEC_FEATURES,
+ .get_features = bcm87xx_get_features,
.config_init = bcm87xx_config_init,
.config_aneg = bcm87xx_config_aneg,
.read_status = bcm87xx_read_status,
@@ -206,7 +203,7 @@ static struct phy_driver bcm87xx_driver[] = {
.phy_id = PHY_ID_BCM8727,
.phy_id_mask = 0xffffffff,
.name = "Broadcom BCM8727",
- .features = PHY_10GBIT_FEC_FEATURES,
+ .get_features = bcm87xx_get_features,
.config_init = bcm87xx_config_init,
.config_aneg = bcm87xx_config_aneg,
.read_status = bcm87xx_read_status,
Hi Heiner
The product briefs for these two devices say:
MDIO interface compliant to IEEE802.3ae Clause 45 with
extended indirect address register access
It might be possible to just detect most of this?
Unfortunately I have no datasheet. There would be more things to
fix in this old driver. The C45 register accesses should be
switched to proper phy_mmd_read/write calls.
Checking the datasheet would not be a problem, I can reply later next
week when I am back to work.
Florian: Do you have access to any boards using these devices?
That I do not unfortunately, those are early 10G PHYs.
--
Florian
