RE: [PATCH net-next v1 0/3] r816e: Add r816e driver in Realtek

Posted by javen 1 week, 2 days ago

>On 1/27/2026 12:34 PM, javen wrote:
>> From: Javen Xu <javen_xu@realsil.com.cn>
>>
>> This series patch add a fake driver r816e to ensure the kernel could 
>> set the correct power state for #fun through r816e driver when suspending.
>>
>
> I think a better approach would be to contribute to adding phylink/sfp support
> to r8169, so that the SFP NIC versions (incl. RTL8116AF) can be properly
> supported. This would have to include making the SFP I2C bus accessible by
> the kernel (instead of firmware only).
>

RTL8116af is a multi-function device. fun0 is a BMC virtual driver and fun1 is
the nic driver. The BMC virtual driver(fun0) is used for power management,
while fun1 provides the actual network functionality.

Thanks,
Javen Xu
>> Javen Xu (3):
>>   r816e: add a fake driver r816e for RTL8116af
>>   r816e: Add a Makefile in the r816e folder
>>   realtek: Update Makefile and Kconfig
>>
>>  drivers/net/ethernet/realtek/Kconfig        |  7 ++
>>  drivers/net/ethernet/realtek/Makefile       |  1 +
>>  drivers/net/ethernet/realtek/r816e/Makefile |  8 +++  
>> drivers/net/ethernet/realtek/r816e/r816e.c  | 75 +++++++++++++++++++++
>>  4 files changed, 91 insertions(+)
>>  create mode 100644 drivers/net/ethernet/realtek/r816e/Makefile
>>  create mode 100644 drivers/net/ethernet/realtek/r816e/r816e.c
>>

Re: [PATCH net-next v1 0/3] r816e: Add r816e driver in Realtek

Posted by Heiner Kallweit 1 week, 2 days ago

On 1/28/2026 7:53 AM, javen wrote:
>> On 1/27/2026 12:34 PM, javen wrote:
> 
>>> From: Javen Xu <javen_xu@realsil.com.cn>
> 
>>>
> 
>>> This series patch add a fake driver r816e to ensure the kernel could 
> 
>>> set the correct power state for #fun through r816e driver when suspending.
> 
>>>
> 
>>
> 
>> I think a better approach would be to contribute to adding phylink/sfp support
> 
>> to r8169, so that the SFP NIC versions (incl. RTL8116AF) can be properly
> 
>> supported. This would have to include making the SFP I2C bus accessible by
> 
>> the kernel (instead of firmware only).
> 
>>
> 
> 
> 
> RTL8116af is a multi-function device. fun0 is a BMC virtual driver and fun1 is
> 
> the nic driver. The BMC virtual driver(fun0) is used for power management,
> 
> while fun1 provides the actual network functionality.
> 
> 
If only purpose of this dummy driver is to bring the device to D3hot during system
suspend, then this may be a question to the PCI subsystem maintainer(s).
PCI devices are instantiated by the PCI core, just that in case of the device here
no driver gets attached. Still PCI core applies a default suspend handling to each
device.

Re: [PATCH net-next v1 0/3] r816e: Add r816e driver in Realtek

Posted by Andrew Lunn 1 week, 2 days ago

> RTL8116af is a multi-function device. fun0 is a BMC virtual driver and fun1 is
> the nic driver. The BMC virtual driver(fun0) is used for power management,
> while fun1 provides the actual network functionality.

This is the sort of thing which should be in patch 0/X, because at the
moment, this patch series makes no sense to me. You need to explain
the big picture, especially when doing something so odd.

What does this BMC virtual driver do? Is it providing an NCSI like
device? Should it have a real driver at some point in the future, more
than this stub?

> >>   r816e: add a fake driver r816e for RTL8116af

Please could you explain the naming. The device is RTL8116af, yet the
files are r816e.c ?

	Andrew