[PATCH 2/3][MTD] P4080/nand: Only make elbc nand driver detect nand flash partitions

[Zang Roy-R61911]

> -----Original Message-----
> From: Anton Vorontsov [mailto:cbouatmailru at gmail.com]
> Sent: Monday, September 06, 2010 16:10 PM
> To: Zang Roy-R61911
> Cc: linux-mtd at lists.infradead.org; Lan Chunhe-B25806; linuxppc-dev at ozlabs.org;
> akpm at linux-foundation.org; Gala Kumar-B11780; Wood Scott-B07421
> Subject: Re: [PATCH 2/3][MTD] P4080/nand: Only make elbc nand driver detect
> nand flash partitions
> 
> On Mon, Sep 06, 2010 at 12:49:17PM +0800, Zang Roy-R61911 wrote:
> > > On Fri, Aug 06, 2010 at 10:51:35AM +0800, Roy Zang wrote:
> > > [...]
> > > >
> > > > +static struct fsl_elbc_fcm_ctrl *elbc_fcm_ctrl;
> > > > +
> > >
> > > Are you sure that you want it as a global var? A bit scary change.
> > >
> > > Oh, you probably don't need it, as you can get it from
> > > fsl_lbc_ctrl_dev->nand?
> > Get it form fsl_lbc_ctrl_dev->nand or assign it to
> > fsl_lbc_ctrl_dev->nand in probe?
> 
> I meant to get it from fsl_lbc_ctrl_dev->nand. I.e. in
> probe() you do: fsl_lbc_ctrl_dev->nand = elbc_fcm_ctrl, so
> you probably don't need the global var.
> 
> (fsl_lbc_ctrl_dev seems to be global as well, duh. Well,
> one variable less in the global name space. But I'd
> probably use lbc_np->data to store the LBC private
> struct).
That makes sense.
> 
> Scott seem to be fine with it as there are probably no
> plans to to add several localbus controllers into the SoCs.
> 
> But, I saw a custom board with two MPC82xx SoCs connected
> together, one as a master (core + peripheral devs), and
> other as a slave (its core was halted, and only slave's
> CPM peripheral devices were used by the master CPU).
> 
> I think it is possible to connect two (or more) SoCs in
> a such way so that two or more LBC controllers would
> be visible for the Linux.
That is an interesting case. Do you have any thought here?
Thanks.
Roy