Skip to contentSkip to conversation

net: dsa: fixed-link interface is reporting SPEED_UNKNOWN

PATCH In review Vladimir Oltean 4 participants 5 messages Apr 11, 2019 – May 6, 2019

Subsystems (3)
lore
Share & tools
lore

Vladimir Oltean

author maintainer
#0 lore raw

PATCH 1/1 net: dsa: fixed-link interface is reporting SPEED_UNKNOWN

With Heiner's recent patch "b6163f194c69 net: phy: improve genphy_read_status", the phydev->speed is now initialized by default to SPEED_UNKNOWN even for fixed PHYs. This is not necessarily bad, since it is not correct to call genphy_config_init() and genphy_read_status() for a fixed PHY.

This dates back all the way to "39b0c705195e net: dsa: Allow configuration of CPU & DSA port speeds/duplex" (discussion thread:
https://www.spinics.net/lists/netdev/msg340862.html).

I don't seem to understand why these calls were necessary back then, but removing these calls seemingly has no impact now apart from preventing the phydev->speed that was set in of_phy_register_fixed_link() from getting overwritten.

---
 net/dsa/port.c | 3 ---
 1 file changed, 3 deletions(-)
net/dsa/port.c +0 -3
index 87769cb38c31..481412c892a7 100644
@@ -485,9 +485,6 @@ static int dsa_port_fixed_link_register_of(struct dsa_port *dp)
 		mode = PHY_INTERFACE_MODE_NA;
 	phydev->interface = mode;
 
-	genphy_config_init(phydev);
-	genphy_read_status(phydev);
-
 	if (ds->ops->adjust_link)
 		ds->ops->adjust_link(ds, port, phydev);
 
Signature · 1 line
2.17.1
#1 lore raw
Quoted 6 lines · Vladimir Oltean

With Heiner's recent patch "b6163f194c69 net: phy: improve genphy_read_status", the phydev->speed is now initialized by default to SPEED_UNKNOWN even for fixed PHYs. This is not necessarily bad, since it is not correct to call genphy_config_init() and genphy_read_status() for a fixed PHY.

What do you mean with "it is not correct"? Whether the calls are always needed may be a valid question, but it's not forbidden to use these calls with a fixed PHY. Actually in phylib polling mode genphy_read_status is called every second also for a fixed PHY. swphy emulates all relevant PHY registers.

Quoted 9 lines

This dates back all the way to "39b0c705195e net: dsa: Allow configuration of CPU & DSA port speeds/duplex" (discussion thread:
https://www.spinics.net/lists/netdev/msg340862.html).

I don't seem to understand why these calls were necessary back then, but removing these calls seemingly has no impact now apart from preventing the phydev->speed that was set in of_phy_register_fixed_link() from getting overwritten.

I tend to agree with the change itself, but not with the justification. For genphy_config_init I'd rather say:
genphy_config_init removes invalid modes based on the abilities read from the chip. But as we emulate all registers anyway, this doesn't provide any benefit for a swphy.

Quoted 20 lines
---
 net/dsa/port.c | 3 ---
 1 file changed, 3 deletions(-)
net/dsa/port.c +0 -3
index 87769cb38c31..481412c892a7 100644
@@ -485,9 +485,6 @@ static int dsa_port_fixed_link_register_of(struct dsa_port *dp)
 		mode = PHY_INTERFACE_MODE_NA;
 	phydev->interface = mode;
 
-	genphy_config_init(phydev);
-	genphy_read_status(phydev);
-
 	if (ds->ops->adjust_link)
 		ds->ops->adjust_link(ds, port, phydev);
 
author maintainer
#2 lore raw

Decoupling phy_device from net_device (was "Re: [PATCH] net: dsa: fixed-link interface is reporting SPEED_UNKNOWN")

Quoted 41 lines · Heiner Kallweit
Quoted 6 lines · Vladimir Oltean

With Heiner's recent patch "b6163f194c69 net: phy: improve genphy_read_status", the phydev->speed is now initialized by default to SPEED_UNKNOWN even for fixed PHYs. This is not necessarily bad, since it is not correct to call genphy_config_init() and genphy_read_status() for a fixed PHY.

What do you mean with "it is not correct"? Whether the calls are always needed may be a valid question, but it's not forbidden to use these calls with a fixed PHY. Actually in phylib polling mode genphy_read_status is called every second also for a fixed PHY. swphy emulates all relevant PHY registers.

Quoted 9 lines

This dates back all the way to "39b0c705195e net: dsa: Allow configuration of CPU & DSA port speeds/duplex" (discussion thread:
https://www.spinics.net/lists/netdev/msg340862.html).

I don't seem to understand why these calls were necessary back then, but removing these calls seemingly has no impact now apart from preventing the phydev->speed that was set in of_phy_register_fixed_link() from getting overwritten.

I tend to agree with the change itself, but not with the justification. For genphy_config_init I'd rather say:
genphy_config_init removes invalid modes based on the abilities read from the chip. But as we emulate all registers anyway, this doesn't provide any benefit for a swphy.

Quoted 13 lines
---
  net/dsa/port.c | 3 ---
  1 file changed, 3 deletions(-)
net/dsa/port.c +0 -0
index 87769cb38c31..481412c892a7 100644
@@ -485,9 +485,6 @@ static int dsa_port_fixed_link_register_of(struct dsa_port *dp)
  		mode = PHY_INTERFACE_MODE_NA;
  	phydev->interface = mode;
 
Quoted 8 lines

net/dsa/port.c static int dsa_port_fixed_link_register_of(struct dsa_port *dp) · line 490

Quoted 7 lines
net/dsa/port.c +0 -3
-	genphy_config_init(phydev);
-	genphy_read_status(phydev);
-
  	if (ds->ops->adjust_link)
  		ds->ops->adjust_link(ds, port, phydev);
 

Hi,

I'd like to resend this, but I'm actually thinking of a nicer way to deal with the whole situation.
Would people be interested in making phylib/phylink decoupled from the phydev->attached_dev? The PHY state machine could be made to simply call a notifier block (similar to how switchdev works) registered by interested parties (MAC driver).
To keep API compatibility (phylink_create), this notifier block could be put inside the net_device structure and the PHY state machine would call it. But a new phylink_create_raw could be added, which passes only the notifier block with no net_device. The CPU port and DSA ports would be immediate and obvious users of this (since they don't register net devices), but there are some other out-of-tree Ethernet drivers out there that have strange workarounds (create a net device that they don't register) to have the PHY driver do its work. Wondering what's your opinion on this and if it would be worth exploring.

Thanks,
-Vladimir

#3 lore raw

Re: Decoupling phy_device from net_device (was "Re: [PATCH] net: dsa: fixed-link interface is reporting SPEED_UNKNOWN")

Quoted 75 lines · Vladimir Oltean
Quoted 3 lines · Heiner Kallweit
Quoted 2 lines · Vladimir Oltean

With Heiner's recent patch "b6163f194c69 net: phy: improve genphy_read_status", the phydev->speed is now initialized by default

to

Quoted 1 line
Quoted 1 line

SPEED_UNKNOWN even for fixed PHYs. This is not necessarily bad,

since it

Quoted 1 line
Quoted 1 line

is not correct to call genphy_config_init() and genphy_read_status()

for

Quoted 3 lines
Quoted 2 lines

a fixed PHY.

What do you mean with "it is not correct"? Whether the calls are

always

Quoted 1 line

needed may be a valid question, but it's not forbidden to use these

calls

Quoted 1 line

with a fixed PHY. Actually in phylib polling mode genphy_read_status

is

Quoted 8 lines

called every second also for a fixed PHY. swphy emulates all relevant PHY registers.

Quoted 5 lines

This dates back all the way to "39b0c705195e net: dsa: Allow configuration of CPU & DSA port speeds/duplex" (discussion thread:
https://www.spinics.net/lists/netdev/msg340862.html).

I don't seem to understand why these calls were necessary back then,

but

Quoted 1 line
Quoted 1 line

removing these calls seemingly has no impact now apart from

preventing

Quoted 4 lines
Quoted 3 lines

the phydev->speed that was set in of_phy_register_fixed_link() from getting overwritten.

I tend to agree with the change itself, but not with the

justification.

Quoted 10 lines

For genphy_config_init I'd rather say:
genphy_config_init removes invalid modes based on the abilities read from the chip. But as we emulate all registers anyway, this doesn't provide any benefit for a swphy.

Quoted 5 lines
---
  net/dsa/port.c | 3 ---
  1 file changed, 3 deletions(-)
Quoted 5 lines

net/dsa/port.c static int

Quoted 5 lines
net/dsa/port.c +0 -0
index 87769cb38c31..481412c892a7 100644
@@ -485,9 +485,6 @@ static int

dsa_port_fixed_link_register_of(struct dsa_port *dp)

Quoted 11 lines

net/dsa/port.c static int dsa_port_fixed_link_register_of(struct dsa_port *dp) · line 490

Quoted 10 lines
net/dsa/port.c +0 -3
  		mode = PHY_INTERFACE_MODE_NA;
  	phydev->interface = mode;
 
-	genphy_config_init(phydev);
-	genphy_read_status(phydev);
-
  	if (ds->ops->adjust_link)
  		ds->ops->adjust_link(ds, port, phydev);
 

Hi,

I'd like to resend this, but I'm actually thinking of a nicer way to deal with the whole situation.
Would people be interested in making phylib/phylink decoupled from the phydev->attached_dev? The PHY state machine could be made to simply call
a notifier block (similar to how switchdev works) registered by interested parties (MAC driver).
To keep API compatibility (phylink_create), this notifier block could be
put inside the net_device structure and the PHY state machine would call
it. But a new phylink_create_raw could be added, which passes only the

Quoted 8 lines

notifier block with no net_device. The CPU port and DSA ports would be immediate and obvious users of this (since they don't register net devices), but there are some other out-of-tree Ethernet drivers out there that have strange workarounds (create a net device that they don't
register) to have the PHY driver do its work.
Wondering what's your opinion on this and if it would be worth exploring.

If you have patches for that idea, I would be glad to take a look. The current way of supporting DSA and CPU ports in DSA is now starting to show its limitations and we are not able to properly make use of PHYLINK at all for those ports. For PHYLIB (outside of PHYLINK), I would not want the decoupling to lead to e.g.: supporting Ethernet switches with only a single net_device instance and managing all the ports using the PHYLIB notifier, because that would bypass the model that DSA offers so we could miss opportunities to fix it, if needed.

Signature · 1 line
Florian
maintainer
#4 lore raw

Re: Decoupling phy_device from net_device (was "Re: [PATCH] net: dsa: fixed-link interface is reporting SPEED_UNKNOWN")

In reply to Vladimir Oltean
Quoted 14 lines · Vladimir Oltean
Quoted 13 lines · Heiner Kallweit
Quoted 6 lines · Vladimir Oltean

With Heiner's recent patch "b6163f194c69 net: phy: improve genphy_read_status", the phydev->speed is now initialized by default to SPEED_UNKNOWN even for fixed PHYs. This is not necessarily bad, since it is not correct to call genphy_config_init() and genphy_read_status() for a fixed PHY.

What do you mean with "it is not correct"? Whether the calls are always needed may be a valid question, but it's not forbidden to use these calls with a fixed PHY. Actually in phylib polling mode genphy_read_status is called every second also for a fixed PHY. swphy emulates all relevant PHY registers.

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines

This dates back all the way to "39b0c705195e net: dsa: Allow configuration of CPU & DSA port speeds/duplex" (discussion thread:
https://www.spinics.net/lists/netdev/msg340862.html).

I don't seem to understand why these calls were necessary back then, but removing these calls seemingly has no impact now apart from preventing the phydev->speed that was set in of_phy_register_fixed_link() from getting overwritten.

As Florian said, if you have patches, please post them and we will consider them.

But i think we also need to take a step back and consider the big picture. There has been a lot of work recently to support multi-G PHYs. It is clear we soon need to make changes to fixed-link. It only supports up to 1G. But we have use cases where we need multi-G fixed links.

We could also consider making the tie to the MAC much stronger. We have been encouraging MAC driver writers to make use of the ndev->phylib pointer. We could even enforce that, and use container_of() to determine the MAC associated to a PHY.

Andrew