Re: [PATCH net] net: dsa: ocelot: add dependency for NET_DSA_MSCC_FELIX

From: maowenan
Date: Tue Nov 19 2019 - 20:28:20 EST




在 2019/11/20 7:41, David Miller 写道:
> From: Mao Wenan <maowenan@xxxxxxxxxx>
> Date: Tue, 19 Nov 2019 10:51:28 +0800
>
>> If CONFIG_NET_DSA_MSCC_FELIX=y, and CONFIG_NET_VENDOR_MICROSEMI=n,
>> below errors can be found:
>> drivers/net/dsa/ocelot/felix.o: In function `felix_vlan_del':
>> felix.c:(.text+0x26e): undefined reference to `ocelot_vlan_del'
>> drivers/net/dsa/ocelot/felix.o: In function `felix_vlan_add':
>> felix.c:(.text+0x352): undefined reference to `ocelot_vlan_add'
>>
>> and warning as below:
>> WARNING: unmet direct dependencies detected for MSCC_OCELOT_SWITCH
>> Depends on [n]: NETDEVICES [=y] && ETHERNET [=y] &&
>> NET_VENDOR_MICROSEMI [=n] && NET_SWITCHDEV [=y] && HAS_IOMEM [=y]
>> Selected by [y]:
>> NET_DSA_MSCC_FELIX [=y] && NETDEVICES [=y] && HAVE_NET_DSA [=y]
>> && NET_DSA [=y] && PCI [=y]
>>
>> This patch add dependency NET_VENDOR_MICROSEMI for NET_DSA_MSCC_FELIX.
>>
>> Fixes: 56051948773e ("net: dsa: ocelot: add driver for Felix switch family")
>> Signed-off-by: Mao Wenan <maowenan@xxxxxxxxxx>
>
> This seems more like a "select" situation, why in the world should the
> user be required to know about NET_VENDOR_MISCROSEMI at all for this
> driver?
thanks.
I will change to 'select' it and send v2.
>
> And NET_VENDOR_MICROSEMI does _NOT_ enable any code at all, you have
> to enable the individual drivers guarded by NET_VENDOR_MICROSEMI in order
> to resolve the symbols necessary for ocelot.
>
> I'm not applying this, it isn't correct.
>
> Thank you.
>
> .
>