Re: [PATCH] ARM: OMAP: AM33XX: CAN: d_can: Add support for Bosch D_CAN controller

[Gary Thomas <gary@mlbassoc.com>]

On 2012-04-05 08:15, AnilKumar, Chimata wrote:
> [dropped some cross-posted lists I had added as requested by Oliver]
>
> On Wed, Apr 04, 2012 at 21:51:41, Wolfgang Grandegger wrote:
>
>> I did look into the manual. Unfortunately, a direct mapping of the C_CAN
>> to the D_CAN registers seems not possible. It's not just a different
>> alignment but sometimes two 16-bit C_CAN registers are folded into *one*
>> 32-bit D_CAN register. Therefore we need something more clever, e.g.
>
> Yes, it's definitely not straightforward.
>
>> using a separate struct or union or handling those register separately.
>> I still think, if feasible, we should avoid an extra driver for the
>> D_CAN controller, also because we sooner than later need the same
>> infrastructure (register_c_can_dev etc.).
>
> Okay, I started hacking the c_can driver to add d_can support
> this way. I will update soon with results/patches.
>
> I hope too that we don't have to add an extra driver.

Sorry to drop in late on this discussion, but what's wrong with
just using the d_can driver from TI?  I have successfully used
it on AM3874 and I'm pretty sure that maps onto AM33xx as well.

Note: as far as I know, this driver is only in the TI PSP kernels.
I extracted it from an Angstrom build for the BeagleBone which
is AM33xx based.

-- 
------------------------------------------------------------
Gary Thomas                 |  Consulting for the
MLB Associates              |    Embedded world
------------------------------------------------------------