[PATCH 2.6.20] powerpc: Changed gianfar device tree definition tomake it more flexible

[Andy Fleming]

On Dec 4, 2006, at 02:20, Joakim Tjernlund wrote:

>
>> +
>> +static struct gfar_flag_info __initdata flag_info[] = {
>> +	{ "gigabit", FSL_GIANFAR_DEV_HAS_GIGABIT },
>> +	{ "coalescing", FSL_GIANFAR_DEV_HAS_COALESCE },
>> +	{ "rmon", FSL_GIANFAR_DEV_HAS_RMON },
>> +	{ "checksumming", FSL_GIANFAR_DEV_HAS_CSUM },
>> +	{ "vlan", FSL_GIANFAR_DEV_HAS_VLAN },
>> +	{ "extended-hash", FSL_GIANFAR_DEV_HAS_EXTENDED_HASH },
>> +	{ "padding", FSL_GIANFAR_DEV_HAS_PADDING },
>> +	{ "filer", FSL_GIANFAR_DEV_HAS_FILER },
>> +	{ "parseL4", FSL_GIANFAR_DEV_HAS_PARSE_L4 },
>> +	{ "parseL3", FSL_GIANFAR_DEV_HAS_PARSE_L3 },
>> +	{ "parseL2", FSL_GIANFAR_DEV_HAS_PARSE_L2 },
>> +	{ "multi-queue", FSL_GIANFAR_DEV_HAS_MULTI_QUEUE },
>> +	{ "buffer-stashing", FSL_GIANFAR_DEV_HAS_BUF_STASHING },
>> +	{ "buffer-locking", FSL_GIANFAR_DEV_HAS_BUF_LOCKING },
>> +	{ "bd-stashing", FSL_GIANFAR_DEV_HAS_BD_STASHING },
>> +	{ "bd-locking", FSL_GIANFAR_DEV_HAS_BD_LOCKING },
>> +};
>
> Shouldn't these be  mapped to generic names(drop the GIANFAR name)?
> The newer UCC also have sevral of these capabilities and it seems
> to me that one can use the same names instead of making a similar  
> table.


That's an excellent idea.  But the UCC code base is still pretty  
foreign to me, so I will have to pursue that with its authors.


Andy