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

[Jonas Gorski]

On Wed, Mar 18, 2015 at 12:09 PM, Rafał Miłecki <zajec5 at gmail.com> wrote:
> 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.

Sorry for taking so long to respond, the kernel-size issue on bcm63xx
took longer and had more side effects than expected.

Then let's keep it simple for now and move it into a separate branch
if you actually have the code ready. At least for me, my future code
never stays that way how I imagined it will be and it tends to look a
lot different when done.


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