[OpenWrt-Devel] [PATCH] b53: fix overriding port 8 state (if it is connected to CPU)

[Rafał Miłecki]

On 18 March 2015 at 11:28, Jonas Gorski <jogo at openwrt.org> wrote:
> On Wed, Mar 18, 2015 at 10:02 AM, Rafał Miłecki <zajec5 at gmail.com> wrote:
>> Signed-off-by: Rafał Miłecki <zajec5 at gmail.com>
>> ---
>>  .../generic/files/drivers/net/phy/b53/b53_common.c | 23 +++++++++++++++++++++-
>>  .../generic/files/drivers/net/phy/b53/b53_regs.h   |  1 +
>>  2 files changed, 23 insertions(+), 1 deletion(-)
>>
>> diff --git a/target/linux/generic/files/drivers/net/phy/b53/b53_common.c b/target/linux/generic/files/drivers/net/phy/b53/b53_common.c
>> index e44d194..4597742 100644
>> --- a/target/linux/generic/files/drivers/net/phy/b53/b53_common.c
>> +++ b/target/linux/generic/files/drivers/net/phy/b53/b53_common.c
>> @@ -525,7 +525,7 @@ static int b53_switch_reset(struct b53_device *dev)
>>                                 return -EINVAL;
>>                         }
>>                 }
>> -       } else if ((is531x5(dev) || is5301x(dev)) && dev->sw_dev.cpu_port == B53_CPU_PORT) {
>> +       } else if (is531x5(dev) && dev->sw_dev.cpu_port == B53_CPU_PORT) {
>>                 u8 mii_port_override;
>>
>>                 b53_read8(dev, B53_CTRL_PAGE, B53_PORT_OVERRIDE_CTRL,
>> @@ -533,6 +533,27 @@ static int b53_switch_reset(struct b53_device *dev)
>>                 b53_write8(dev, B53_CTRL_PAGE, B53_PORT_OVERRIDE_CTRL,
>>                            mii_port_override | PORT_OVERRIDE_EN |
>>                            PORT_OVERRIDE_LINK);
>> +       } else if (is5301x(dev)) {
>> +               /*
>> +                * CPU interface attached to port 8 requires specific handling.
>> +                * It uses different overriding register and extra ports 5 and 7
>> +                * need to be configured as well.
>> +                */
>> +               if (dev->sw_dev.cpu_port == 8) {
>> +                       u8 mii_port_override;
>> +
>> +                       b53_read8(dev, B53_CTRL_PAGE, B53_PORT_OVERRIDE_CTRL,
>> +                                 &mii_port_override);
>> +                       mii_port_override |= PORT_OVERRIDE_LINK |
>> +                                            PORT_OVERRIDE_RX_FLOW |
>> +                                            PORT_OVERRIDE_TX_FLOW |
>> +                                            PORT_OVERRIDE_SPEED_2000M |
>> +                                            PORT_OVERRIDE_EN;
>> +                       b53_write8(dev, B53_CTRL_PAGE, B53_PORT_OVERRIDE_CTRL,
>> +                                  mii_port_override);
>> +               } else {
>> +                       pr_warn("overriding CPU port other than 8 is not supported yet\n");
>> +               }
>>         }
>>
>>         b53_enable_mib(dev);
>
> How about
>
> @@ -530,9 +530,16 @@ static int b53_switch_reset(struct b53_device *dev)
>
>                 b53_read8(dev, B53_CTRL_PAGE, B53_PORT_OVERRIDE_CTRL,
>                           &mii_port_override);
> +
> +               mii_port_override |= PORT_OVERRIDE_LINK | PORT_OVERRIDE_EN;
> +
> +               if (is5301x(dev))
> +                       mii_port_override |= PORT_OVERRIDE_RX_FLOW |
> +                                            PORT_OVERRIDE_TX_FLOW |
> +                                            PORT_OVERRIDE_SPEED_2000M;
> +
>                 b53_write8(dev, B53_CTRL_PAGE, B53_PORT_OVERRIDE_CTRL,
> -                          mii_port_override | PORT_OVERRIDE_EN |
> -                          PORT_OVERRIDE_LINK);
> +                          mii_port_override);
>         }
>
>         b53_enable_mib(dev);
>
> instead of creating a full new branch?

As the comment says, this code for BCM5301X will be extended. This is
because of this 2000M speed, which requires configuring 3 ports. It
seems communication between switch and CPU interface with such speed
couldn't be handled with only a single port. Broadcom decided to use 3
ports in total for that.

I don't have that code ready and I also don't like patch bombs, so I
started with this simple change. However because of further
development plans I vote for a separated branch.

-- 
Rafał
_______________________________________________
openwrt-devel mailing list
openwrt-devel at lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel