diff mbox

ARM: dts: armada388-clearfog: number LAN ports properly

Message ID E1bLWIh-0001ZK-Ai@rmk-PC.armlinux.org.uk (mailing list archive)
State New, archived
Headers show

Commit Message

Russell King (Oracle) July 8, 2016, 1:58 p.m. UTC
Currently, the ports as seen from the rear number as:

	eth0 sfp lan5 lan4 lan3 lan2 lan1 lan6

which is illogical - this came about because the rev 2.0 boards have the
LEDs on the front for the DSA switch (lan5-1) reversed.  Rev 2.1 boards
fixed the LED issue, and the Clearfog case numbers the lan ports
increasing from left to right.

Maintaining this illogical numbering causes confusion, with reports that
"my link isn't coming up" and "my connection negotiates 10base-Half"
both of which are due to people thinking that the port next to the SFP
is lan1.

Fix this by renumbering the ports to match people's expectations.

Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
---
 arch/arm/boot/dts/armada-388-clearfog.dts | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

Comments

Andrew Lunn July 8, 2016, 3:26 p.m. UTC | #1
On Fri, Jul 08, 2016 at 02:58:39PM +0100, Russell King wrote:
> Currently, the ports as seen from the rear number as:
> 
> 	eth0 sfp lan5 lan4 lan3 lan2 lan1 lan6
> 
> which is illogical - this came about because the rev 2.0 boards have the
> LEDs on the front for the DSA switch (lan5-1) reversed.  Rev 2.1 boards
> fixed the LED issue, and the Clearfog case numbers the lan ports
> increasing from left to right.
> 
> Maintaining this illogical numbering causes confusion, with reports that
> "my link isn't coming up" and "my connection negotiates 10base-Half"
> both of which are due to people thinking that the port next to the SFP
> is lan1.
> 
> Fix this by renumbering the ports to match people's expectations.
> 
> Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>

Reviewed-by: Andrew Lunn <andrew@lunn.ch>

    Andrew
Gregory CLEMENT July 27, 2016, 10:19 a.m. UTC | #2
Hi,
 
 On ven., juil. 08 2016, Andrew Lunn <andrew@lunn.ch> wrote:

> On Fri, Jul 08, 2016 at 02:58:39PM +0100, Russell King wrote:
>> Currently, the ports as seen from the rear number as:
>> 
>> 	eth0 sfp lan5 lan4 lan3 lan2 lan1 lan6
>> 
>> which is illogical - this came about because the rev 2.0 boards have the
>> LEDs on the front for the DSA switch (lan5-1) reversed.  Rev 2.1 boards
>> fixed the LED issue, and the Clearfog case numbers the lan ports
>> increasing from left to right.
>> 
>> Maintaining this illogical numbering causes confusion, with reports that
>> "my link isn't coming up" and "my connection negotiates 10base-Half"
>> both of which are due to people thinking that the port next to the SFP
>> is lan1.
>> 
>> Fix this by renumbering the ports to match people's expectations.
>> 
>> Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
>
> Reviewed-by: Andrew Lunn <andrew@lunn.ch>

I missed this patch, but I now applied it on mvebu/dt-4.9

Thanks,

Gregory
Russell King (Oracle) July 27, 2016, 10:21 a.m. UTC | #3
On Wed, Jul 27, 2016 at 12:19:01PM +0200, Gregory CLEMENT wrote:
> Hi,
>  
>  On ven., juil. 08 2016, Andrew Lunn <andrew@lunn.ch> wrote:
> 
> > On Fri, Jul 08, 2016 at 02:58:39PM +0100, Russell King wrote:
> >> Currently, the ports as seen from the rear number as:
> >> 
> >> 	eth0 sfp lan5 lan4 lan3 lan2 lan1 lan6
> >> 
> >> which is illogical - this came about because the rev 2.0 boards have the
> >> LEDs on the front for the DSA switch (lan5-1) reversed.  Rev 2.1 boards
> >> fixed the LED issue, and the Clearfog case numbers the lan ports
> >> increasing from left to right.
> >> 
> >> Maintaining this illogical numbering causes confusion, with reports that
> >> "my link isn't coming up" and "my connection negotiates 10base-Half"
> >> both of which are due to people thinking that the port next to the SFP
> >> is lan1.
> >> 
> >> Fix this by renumbering the ports to match people's expectations.
> >> 
> >> Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
> >
> > Reviewed-by: Andrew Lunn <andrew@lunn.ch>
> 
> I missed this patch, but I now applied it on mvebu/dt-4.9

It would be much better to get it into 4.8-rc so that we don't spread
the port renumbering over a large range of kernels, as well as forcing
vendors to carry patches like this to fix problems.
Gregory CLEMENT July 27, 2016, 10:26 a.m. UTC | #4
Hi Russell King,
 
 On mer., juil. 27 2016, Russell King - ARM Linux <linux@armlinux.org.uk> wrote:

> On Wed, Jul 27, 2016 at 12:19:01PM +0200, Gregory CLEMENT wrote:
>> Hi,
>>  
>>  On ven., juil. 08 2016, Andrew Lunn <andrew@lunn.ch> wrote:
>> 
>> > On Fri, Jul 08, 2016 at 02:58:39PM +0100, Russell King wrote:
>> >> Currently, the ports as seen from the rear number as:
>> >> 
>> >> 	eth0 sfp lan5 lan4 lan3 lan2 lan1 lan6
>> >> 
>> >> which is illogical - this came about because the rev 2.0 boards have the
>> >> LEDs on the front for the DSA switch (lan5-1) reversed.  Rev 2.1 boards
>> >> fixed the LED issue, and the Clearfog case numbers the lan ports
>> >> increasing from left to right.
>> >> 
>> >> Maintaining this illogical numbering causes confusion, with reports that
>> >> "my link isn't coming up" and "my connection negotiates 10base-Half"
>> >> both of which are due to people thinking that the port next to the SFP
>> >> is lan1.
>> >> 
>> >> Fix this by renumbering the ports to match people's expectations.
>> >> 
>> >> Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
>> >
>> > Reviewed-by: Andrew Lunn <andrew@lunn.ch>
>> 
>> I missed this patch, but I now applied it on mvebu/dt-4.9
>
> It would be much better to get it into 4.8-rc so that we don't spread
> the port renumbering over a large range of kernels, as well as forcing
> vendors to carry patches like this to fix problems.

I can move it to the mvebu/fixes branch, it is not too late. Also what
about to apply it on the stable kernel?

Gregory

>
> -- 
> RMK's Patch system: http://www.armlinux.org.uk/developer/patches/
> FTTC broadband for 0.8mile line: currently at 9.6Mbps down 400kbps up
> according to speedtest.net.
Russell King (Oracle) July 27, 2016, 10:42 a.m. UTC | #5
On Wed, Jul 27, 2016 at 12:26:41PM +0200, Gregory CLEMENT wrote:
> Hi Russell King,
>  
>  On mer., juil. 27 2016, Russell King - ARM Linux <linux@armlinux.org.uk> wrote:
> 
> > On Wed, Jul 27, 2016 at 12:19:01PM +0200, Gregory CLEMENT wrote:
> >> Hi,
> >>  
> >>  On ven., juil. 08 2016, Andrew Lunn <andrew@lunn.ch> wrote:
> >> 
> >> > On Fri, Jul 08, 2016 at 02:58:39PM +0100, Russell King wrote:
> >> >> Currently, the ports as seen from the rear number as:
> >> >> 
> >> >> 	eth0 sfp lan5 lan4 lan3 lan2 lan1 lan6
> >> >> 
> >> >> which is illogical - this came about because the rev 2.0 boards have the
> >> >> LEDs on the front for the DSA switch (lan5-1) reversed.  Rev 2.1 boards
> >> >> fixed the LED issue, and the Clearfog case numbers the lan ports
> >> >> increasing from left to right.
> >> >> 
> >> >> Maintaining this illogical numbering causes confusion, with reports that
> >> >> "my link isn't coming up" and "my connection negotiates 10base-Half"
> >> >> both of which are due to people thinking that the port next to the SFP
> >> >> is lan1.
> >> >> 
> >> >> Fix this by renumbering the ports to match people's expectations.
> >> >> 
> >> >> Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
> >> >
> >> > Reviewed-by: Andrew Lunn <andrew@lunn.ch>
> >> 
> >> I missed this patch, but I now applied it on mvebu/dt-4.9
> >
> > It would be much better to get it into 4.8-rc so that we don't spread
> > the port renumbering over a large range of kernels, as well as forcing
> > vendors to carry patches like this to fix problems.
> 
> I can move it to the mvebu/fixes branch, it is not too late. Also what
> about to apply it on the stable kernel?

That'd probably be best.

As far as stable goes, I can't convince myself that it is really stable
kernel material.  I'm not aware what happened to the eth* renumber patch,
was that applied to stable trees?  My view would be that the lan*
renumbering should be applied to the same trees which include the eth*
renumbering and no further, iff the eth* renumber was even backported.

Talking to Jon Nettleton @ SR, he's in favour of it being applied to
stable kernels.
diff mbox

Patch

diff --git a/arch/arm/boot/dts/armada-388-clearfog.dts b/arch/arm/boot/dts/armada-388-clearfog.dts
index 54bf0749e15e..42d760da4e82 100644
--- a/arch/arm/boot/dts/armada-388-clearfog.dts
+++ b/arch/arm/boot/dts/armada-388-clearfog.dts
@@ -386,12 +386,12 @@ 
 
 			port@0 {
 				reg = <0>;
-				label = "lan1";
+				label = "lan5";
 			};
 
 			port@1 {
 				reg = <1>;
-				label = "lan2";
+				label = "lan4";
 			};
 
 			port@2 {
@@ -401,12 +401,12 @@ 
 
 			port@3 {
 				reg = <3>;
-				label = "lan4";
+				label = "lan2";
 			};
 
 			port@4 {
 				reg = <4>;
-				label = "lan5";
+				label = "lan1";
 			};
 
 			port@5 {