Skip to contentSkip to conversation

devlink: Expose external flag for PCI SF ports

PATCH Applied Shay Drory 4 participants 4 messages Aug 3 – Aug 5

Subsystems (3)
lore
Share & tools
lore

Shay Drory

author
#0 lore raw

PATCH 1/1 devlink: Expose external flag for PCI SF ports

The external flag is part of the PCI SF port attributes, but unlike the PCI PF and PCI VF flavours it was never filled into the port dump, so userspace could not query it directly.

Reporting of the external flag was missed for SF ports. Hence, put DEVLINK_ATTR_PORT_EXTERNAL for the PCI SF flavour as well, matching what PCI PF and PCI VF ports already report.

$ devlink port show pci/0033:01:00.0/163840
 pci/0033:01:00.0/163840: type eth netdev eth1 flavour pcisf controller 1 pfnum 0 sfnum 77 external true splittable false
Fixes a1ab3e4554b5 ("devlink: Extend SF port attributes to have external attribute")
---
 net/devlink/port.c | 2 ++
 1 file changed, 2 insertions(+)
net/devlink/port.c +2 -0
index dc82cac68e7d..1528f2d148df 100644
@@ -267,6 +267,8 @@ static int devlink_nl_port_attrs_put(struct sk_buff *msg,
 		    nla_put_u32(msg, DEVLINK_ATTR_PORT_PCI_SF_NUMBER,
 				attrs->pci_sf.sf))
 			return -EMSGSIZE;
+		if (nla_put_u8(msg, DEVLINK_ATTR_PORT_EXTERNAL, attrs->pci_sf.external))
+			return -EMSGSIZE;
 		break;
 	case DEVLINK_PORT_FLAVOUR_PHYSICAL:
 	case DEVLINK_PORT_FLAVOUR_CPU:
Signature · 1 line
2.38.1
#1 lore raw
Quoted 4 lines · Shay Drory

The external flag is part of the PCI SF port attributes, but unlike the PCI PF and PCI VF flavours it was never filled into the port dump, so userspace could not query it directly.

Quoted 8 lines

Reporting of the external flag was missed for SF ports. Hence, put DEVLINK_ATTR_PORT_EXTERNAL for the PCI SF flavour as well, matching what PCI PF and PCI VF ports already report.

$ devlink port show pci/0033:01:00.0/163840
  pci/0033:01:00.0/163840: type eth netdev eth1 flavour pcisf controller 1 pfnum 0 sfnum 77 external true splittable false
Fixes a1ab3e4554b5 ("devlink: Extend SF port attributes to have external attribute")

Cited commit has neither introduced a user-observed bug nor a buggy feature, merely a "backend for a feature", and now you exposes it, making "the frontend".
IOW: I will rather call it a next material.

OTOH, code is fine

Quoted 19 lines
---
  net/devlink/port.c | 2 ++
  1 file changed, 2 insertions(+)
net/devlink/port.c +2 -0
index dc82cac68e7d..1528f2d148df 100644
@@ -267,6 +267,8 @@ static int devlink_nl_port_attrs_put(struct sk_buff *msg,
  		    nla_put_u32(msg, DEVLINK_ATTR_PORT_PCI_SF_NUMBER,
  				attrs->pci_sf.sf))
  			return -EMSGSIZE;
+		if (nla_put_u8(msg, DEVLINK_ATTR_PORT_EXTERNAL, attrs->pci_sf.external))
+			return -EMSGSIZE;
  		break;
  	case DEVLINK_PORT_FLAVOUR_PHYSICAL:
  	case DEVLINK_PORT_FLAVOUR_CPU:
#2 lore raw
Quoted 17 lines

From: Przemek Kitszel <[email protected]>
Sent: 03 August 2026 07:20 PM

Quoted 11 lines · Shay Drory

The external flag is part of the PCI SF port attributes, but unlike the PCI PF and PCI VF flavours it was never filled into the port dump, so userspace could not query it directly.

Reporting of the external flag was missed for SF ports. Hence, put DEVLINK_ATTR_PORT_EXTERNAL for the PCI SF flavour as well, matching what PCI PF and PCI VF ports already report.

$ devlink port show pci/0033:01:00.0/163840
  pci/0033:01:00.0/163840: type eth netdev eth1 flavour pcisf controller 1 pfnum 0 sfnum 77 external true splittable false
Quoted 7 lines
Quoted 1 line
Fixes a1ab3e4554b5 ("devlink: Extend SF port attributes to have external attribute")

Cited commit has neither introduced a user-observed bug nor a buggy feature, merely a "backend for a feature", and now you exposes it, making "the frontend".
IOW: I will rather call it a next material.

It is indeed a user observable bug to an extent that SF port is external, but user is not able to see the same. External attribute was exposed as devlink port 'frontend' attribute.

A script that was supposed to work for the external PF, fails to find the attribute on the SF of the external PF.

The cited patch missed to expose the attribute, which was the patch's original intent; which in my view is a bug and fixes tag.

But at the same time, one can also say, that external flag itself was not exposed...

Inclined for net because of presence of fixes tag. Not very strong opinion though...

Quoted 21 lines

OTOH, code is fine

Quoted 19 lines
---
  net/devlink/port.c | 2 ++
  1 file changed, 2 insertions(+)
net/devlink/port.c +2 -0
index dc82cac68e7d..1528f2d148df 100644
@@ -267,6 +267,8 @@ static int devlink_nl_port_attrs_put(struct sk_buff *msg,
  		    nla_put_u32(msg, DEVLINK_ATTR_PORT_PCI_SF_NUMBER,
  				attrs->pci_sf.sf))
  			return -EMSGSIZE;
+		if (nla_put_u8(msg, DEVLINK_ATTR_PORT_EXTERNAL, attrs->pci_sf.external))
+			return -EMSGSIZE;
  		break;
  	case DEVLINK_PORT_FLAVOUR_PHYSICAL:
  	case DEVLINK_PORT_FLAVOUR_CPU:
#3 lore raw
In reply to Shay Drory
This patch was applied to netdev/net-next.git (main) by Jakub Kicinski <[email protected]>:

Hello:

This patch was applied to netdev/net-next.git (main) by Jakub Kicinski <[email protected]>:

Quoted 9 lines · On Mon, 3 Aug 2026 12:00:12 +0300 you wrote:

The external flag is part of the PCI SF port attributes, but unlike the PCI PF and PCI VF flavours it was never filled into the port dump, so userspace could not query it directly.

Reporting of the external flag was missed for SF ports. Hence, put DEVLINK_ATTR_PORT_EXTERNAL for the PCI SF flavour as well, matching what PCI PF and PCI VF ports already report.

[...]

Here is the summary with links:
- [net] devlink: Expose external flag for PCI SF ports

    https://git.kernel.org/netdev/net-next/c/cb59bfd419d0

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html