[PATCH] powerpc: consolidate mpc83xx platform files

Kumar Gala galak at kernel.crashing.org
Mon Dec 11 14:41:24 EST 2006


On Dec 9, 2006, at 1:14 AM, Benjamin Herrenschmidt wrote:

> On Fri, 2006-12-08 at 19:07 -0600, Kim Phillips wrote:
>> Eliminate code redundancy.  mpc83[246]x_{mds,itx,sys,pb} files merged
>> into a single setup.c.  machine_probe, instead of using the model  
>> property,
>> checks the compatible property for "MPC83xx" (dts files updated  
>> appropriately).
>> This patch also utilizes of_platform_bus_probe() in lieu of manually
>> calling of_platform_device_create for each ucc_geth device.
>>
>> Signed-off-by: Kim Phillips <kim.phillips at freescale.com>

Nack on the patch.

> I am not completely certain this is the right approach.
>
> While factoring code is good, I think that every single board should
> have it's own ppc_md, though you can definitely provide "common"
> functions for mpc83xx that can optinally be used by those different
> boards.
>
> Maybe put all the freescale ones in one file if you want...
>
> The rationale here is that while your approach is fine for your eval
> boards, I don't think it's good for embedded customers. They may want
> more complex platforms, with their own directory even if they have  
> a lot
> of custom stuff on the board while still possibly picking some of your
> "common" code (and their board shouldn't match your overly generic
> probe() implementation).
>
> Cheers,
> Ben.

I'm with Ben on this.  I think consolidating the code that is common  
is fine, but we should have a define_machine() per board.  You can  
put them all in one mpc83xx/fsl.c

- k 



More information about the Linuxppc-dev mailing list