Skip to contentSkip to conversation

Ethernet PHY Power Down

U-Boot Michael Welling 2 participants 6 messages Nov 6, 2015

lore
Share & tools
lore
#0 lore raw

Lets try this again with list membership.

I have run into an issue with a KSZ9031 Gigabit Ethernet PHY on an IMX6 module.

After booting into Linux and rebooting the ethernet fails consistently.

With a bit of investigation I found that the power-down bit in the PHY control register is being set on reboot.

=> dhcp
BOOTP broadcast 1
BOOTP broadcast 2
BOOTP broadcast 3
BOOTP broadcast 4
BOOTP broadcast 5
.
.
Abort
=> mii dump 1 0
0. (1940) -- PHY control register --

  (8000:0000) 0.15    =     0    reset
  (4000:0000) 0.14    =     0    loopback
  (2040:0040) 0. 6,13 =   b10    speed selection = 1000 Mbps
  (1000:1000) 0.12    =     1    A/N enable
  (0800:0800) 0.11    =     1    power-down
  (0400:0000) 0.10    =     0    isolate
  (0200:0000) 0. 9    =     0    restart A/N
  (0100:0100) 0. 8    =     1    duplex = full
  (0080:0000) 0. 7    =     0    collision test enable
  (003f:0000) 0. 5- 0 =     0    (reserved)

When I manually toggle the power down bit it starts to work again.

=> mii write 1 0 1140
=> dhcp
FEC Waiting for PHY auto negotiation to complete........ done BOOTP broadcast 1
DHCP client bound to address 192.168.0.12 (62 ms)

I tried performing the above steps in the PHY driver code with little luck.

When I toggle the bit in the code the first dhcp always fails but then it works on subsequent attempts.

Any ideas?
Has anyone else seen this issue with other PHYs?

#1 lore raw
Quoted 39 lines · Michael Welling

Lets try this again with list membership.

I have run into an issue with a KSZ9031 Gigabit Ethernet PHY on an IMX6 module.

After booting into Linux and rebooting the ethernet fails consistently.

With a bit of investigation I found that the power-down bit in the PHY control register is being set on reboot.

=> dhcp
BOOTP broadcast 1
BOOTP broadcast 2
BOOTP broadcast 3
BOOTP broadcast 4
BOOTP broadcast 5
.
.
Abort
=> mii dump 1 0
0. (1940) -- PHY control register --

  (8000:0000) 0.15    =     0    reset
  (4000:0000) 0.14    =     0    loopback
  (2040:0040) 0. 6,13 =   b10    speed selection = 1000 Mbps
  (1000:1000) 0.12    =     1    A/N enable
  (0800:0800) 0.11    =     1    power-down
  (0400:0000) 0.10    =     0    isolate
  (0200:0000) 0. 9    =     0    restart A/N
  (0100:0100) 0. 8    =     1    duplex = full
  (0080:0000) 0. 7    =     0    collision test enable
  (003f:0000) 0. 5- 0 =     0    (reserved)

When I manually toggle the power down bit it starts to work again.

=> mii write 1 0 1140
=> dhcp
FEC Waiting for PHY auto negotiation to complete........ done BOOTP broadcast 1
DHCP client bound to address 192.168.0.12 (62 ms)

Quoted 7 lines

I tried performing the above steps in the PHY driver code with little luck.

When I toggle the bit in the code the first dhcp always fails but then it works on subsequent attempts.

Any ideas?
Has anyone else seen this issue with other PHYs?

Could you try to revert 59370f3fc as suggested here?
https://www.mail-archive.com/u-boot at lists.denx.de/msg191196.html

#2 lore raw
Quoted 42 lines · Fabio Estevam
Quoted 41 lines · Michael Welling

Lets try this again with list membership.

I have run into an issue with a KSZ9031 Gigabit Ethernet PHY on an IMX6 module.

After booting into Linux and rebooting the ethernet fails consistently.

With a bit of investigation I found that the power-down bit in the PHY control register is being set on reboot.

=> dhcp
BOOTP broadcast 1
BOOTP broadcast 2
BOOTP broadcast 3
BOOTP broadcast 4
BOOTP broadcast 5
.
.
Abort
=> mii dump 1 0
0. (1940) -- PHY control register --

  (8000:0000) 0.15    =     0    reset
  (4000:0000) 0.14    =     0    loopback
  (2040:0040) 0. 6,13 =   b10    speed selection = 1000 Mbps
  (1000:1000) 0.12    =     1    A/N enable
  (0800:0800) 0.11    =     1    power-down
  (0400:0000) 0.10    =     0    isolate
  (0200:0000) 0. 9    =     0    restart A/N
  (0100:0100) 0. 8    =     1    duplex = full
  (0080:0000) 0. 7    =     0    collision test enable
  (003f:0000) 0. 5- 0 =     0    (reserved)

When I manually toggle the power down bit it starts to work again.

=> mii write 1 0 1140
=> dhcp
FEC Waiting for PHY auto negotiation to complete........ done BOOTP broadcast 1
DHCP client bound to address 192.168.0.12 (62 ms)

I tried performing the above steps in the PHY driver code with little luck.

Quoted 8 lines
Quoted 5 lines

When I toggle the bit in the code the first dhcp always fails but then it works on subsequent attempts.

Any ideas?
Has anyone else seen this issue with other PHYs?

Could you try to revert 59370f3fc as suggested here?
https://www.mail-archive.com/u-boot at lists.denx.de/msg191196.html

Reverting this does not effect the behavior. It appears that other boards are registering a GPIO to toggle the reset. On this particular board we are using the system reset.

Pressing the reset button does bring the PHY out of power down.

Perhaps I can add the code into our board_phy_config function but this is a generic PHY feature and could effect other systems.

#3 lore raw
In reply to Fabio Estevam
Quoted 42 lines · Fabio Estevam
Quoted 41 lines · Michael Welling

Lets try this again with list membership.

I have run into an issue with a KSZ9031 Gigabit Ethernet PHY on an IMX6 module.

After booting into Linux and rebooting the ethernet fails consistently.

With a bit of investigation I found that the power-down bit in the PHY control register is being set on reboot.

=> dhcp
BOOTP broadcast 1
BOOTP broadcast 2
BOOTP broadcast 3
BOOTP broadcast 4
BOOTP broadcast 5
.
.
Abort
=> mii dump 1 0
0. (1940) -- PHY control register --

  (8000:0000) 0.15    =     0    reset
  (4000:0000) 0.14    =     0    loopback
  (2040:0040) 0. 6,13 =   b10    speed selection = 1000 Mbps
  (1000:1000) 0.12    =     1    A/N enable
  (0800:0800) 0.11    =     1    power-down
  (0400:0000) 0.10    =     0    isolate
  (0200:0000) 0. 9    =     0    restart A/N
  (0100:0100) 0. 8    =     1    duplex = full
  (0080:0000) 0. 7    =     0    collision test enable
  (003f:0000) 0. 5- 0 =     0    (reserved)

When I manually toggle the power down bit it starts to work again.

=> mii write 1 0 1140
=> dhcp
FEC Waiting for PHY auto negotiation to complete........ done BOOTP broadcast 1
DHCP client bound to address 192.168.0.12 (62 ms)

I tried performing the above steps in the PHY driver code with little luck.

Quoted 8 lines
Quoted 5 lines

When I toggle the bit in the code the first dhcp always fails but then it works on subsequent attempts.

Any ideas?
Has anyone else seen this issue with other PHYs?

Could you try to revert 59370f3fc as suggested here?
https://www.mail-archive.com/u-boot at lists.denx.de/msg191196.html

So I added the following lines to my board_phy_config function and the problem seems to be fixed:
.
.

	ctl = phy_read(phydev, MDIO_DEVAD_NONE, MII_BMCR);                                                                                                                                                                                                                   
	if (ctl & BMCR_PDOWN) {                                                                                                                                                                                                                                              
		ctl &= ~BMCR_PDOWN;                                                                                                                                                                                                                                          
		phy_write(phydev, MDIO_DEVAD_NONE, MII_BMCR, ctl);                                                                                                                                                                                                           
		mdelay(15);                                                                                                                                                                                                                                                  
	}

.
.

Should this be handled somewhere in the PHY driver core?

#4 lore raw
Quoted 10 lines · Michael Welling

So I added the following lines to my board_phy_config function and the problem seems to be fixed:
.
.

        ctl = phy_read(phydev, MDIO_DEVAD_NONE, MII_BMCR);
        if (ctl & BMCR_PDOWN) {
                ctl &= ~BMCR_PDOWN;
                phy_write(phydev, MDIO_DEVAD_NONE, MII_BMCR, ctl);
                mdelay(15);
        }
Quoted 5 lines

.
.

Should this be handled somewhere in the PHY driver core?

Adding Joe for his comments.

Regards,

Fabio Estevam

#5 lore raw
Quoted 3 lines · Fabio Estevam
Quoted 2 lines · Michael Welling

So I added the following lines to my board_phy_config function and the problem seems to be fixed:

It appears that I spoke to soon. Occassionally the autonegiotation still times out. This may be a separate issue. I will keep everyone updated.

Quoted 19 lines
Quoted 8 lines

.
.

        ctl = phy_read(phydev, MDIO_DEVAD_NONE, MII_BMCR);
        if (ctl & BMCR_PDOWN) {
                ctl &= ~BMCR_PDOWN;
                phy_write(phydev, MDIO_DEVAD_NONE, MII_BMCR, ctl);
                mdelay(15);
        }
Quoted 5 lines

.
.

Should this be handled somewhere in the PHY driver core?

Adding Joe for his comments.

Regards,

Fabio Estevam